JP6679281B2 - Method and system for enhancing control of a power plant power generation unit - Google Patents

Description

  The invention of the present application relates generally to power generation, and more specifically to methods and systems associated with optimizing and / or enhancing the economics and performance of power plants having thermal power generation units.

  In a power system, a large number of participants or power plants generate electricity, which is then distributed to residential and business customers through common transmission lines. As will be appreciated, thermal power generation units, such as gas turbines, steam turbines, and combined cycle plants, are still utilized to generate a substantial portion of the power required by such systems. Has been done. Each of the power plants in such a system includes one or more power generation units, each of these units typically including a control system to control operation, and also two or more power generation units. In the case of a power plant with units, it includes a control system that controls the performance of the entire power plant. As an example, one of the responsibilities of a plant operator is the generation of an offer curve that represents the cost of electricity production. The offer curve typically includes an incremental variable cost curve, an average variable cost curve, or another suitable indication of variable generation costs, which is typically expressed in dollars per megawatt hour for megawatt output. To be done. The average variable cost curve can represent the cumulative cost divided by the cumulative power output for a given point, and the incremental variable cost curve represents the change in cost divided by the change in power output. It will be appreciated that it is possible. The incremental variable cost curve can be obtained, for example, by taking the first derivative of the input-output curve of the power plant, which represents the cost per hour for the generated power. In combined cycle power plants, where waste heat from a fuel-burning generator is used to create steam for powering an auxiliary steam turbine, an incremental variable cost curve is also obtained using known techniques. Is possible, but its derivation can be more complicated.

  In most power systems, a competitive process, commonly referred to as economical power distribution, is used to divide the system load between power plants over future periods. As part of this process, the power plant periodically generates an offer curve and sends it to the power system authority or dispatch center. Such an offer curve represents a bid from a power plant to generate a portion of the electricity needed by the power system over future market periods. The dispatching authority should receive offer curves from the power plants in its system, evaluate those offer curves and engage each power plant to most efficiently meet the anticipated load requirements of the system. Determine the level. In doing so, the dispatching authority develops a shutdown schedule that describes the limits to which each of the power plants will be involved over the relevant period in order to analyze the offer curve and find the lowest cost of electricity for the system. produce.

  Once the start-up plan schedule is communicated to the power plants, each power plant can determine the most efficient and cost-effective way to satisfy its load start-up plan. It will be appreciated that the power generation unit of a power plant includes a control system that monitors and controls operation. When the power generation unit comprises a thermal power generator, such control system governs the combustion system and other aspects of operation. (For purposes of example, both gas turbines and combined cycle power plants are described herein. However, certain embodiments of the invention may be applied to other types of power generation units, or others It will be appreciated that the control system may be used with any type of power generation unit.) The control system provides fuel flow, inlet guide vanes, and other control inputs to ensure efficient operation of the engine. It is possible to implement a scheduling scheduling algorithm. However, the actual power output and efficiency of a power plant are affected by external factors such as variable ambient conditions that cannot be completely assumed. As will be appreciated, the complexity of such systems, and the variability of operating conditions, make performance difficult to predict and control, which often results in inefficient operation.

  Mechanical degradation that occurs over time is another problem that is difficult to quantify facts, which can have a significant impact on the performance of a power generation unit. The rate of deterioration, the replacement of worn components, the timing of maintenance routines, and other factors affect the short-term performance of the plant, and therefore when generating the cost curve during the feeding process, as well as the plant. It will be recognized that it needs to be taken into account when assessing the long-term cost effectiveness of a. As an example, the life of a gas turbine typically includes limits expressed in both time of operation and number of starts. If the gas turbine or its components reach their start-up limit prior to their time limit, then the gas turbine or its components must be repaired or replaced, even if their time-based lifetime remains. Time-based life in a gas turbine can be extended by reducing the combustion temperature, which reduces the efficiency of the gas turbine, which increases the cost of operation. Conversely, increasing the combustion temperature increases efficiency, but shortens gas turbine life and increases maintenance and / or replacement costs. As will be appreciated, the life cycle cost of a heat engine depends on many complex factors while at the same time representing an important consideration in the economic efficiency of a power plant.

  Given the complexity of modern power plants, especially those with multiple power generation units, and the markets in which they compete, power plant operators continue to struggle to maximize economic benefits. Has continued to do. For example, grid compliance and feed planning for power plants uses static control profiles such as heat rate curves that are derived and gathered only from periodic performance tests in an overly static manner. Then, by controlling the thermal power generation unit, it is adversely affected. During these periodic updates, turbine engine performance may change (eg, from degradation), which may affect start-up and load performance. Moreover, changes in external factors that occur during the day can lead to inefficient operation if this is not taken into account in the turbine control profile. To compensate for this type of variability, power plant operators are often overly conservative when planning for future operation, which results in underutilized power generation units. At other times, the plant operator may be forced to operate the unit inefficiently in order to satisfy the excessive start and stop schedule.

  Without identifying short-term inefficiencies and / or long-term degradation as each is realized, conventional control systems in power plants must be retuned, which is an expensive process. Or, it must be operated conservatively to proactively address component degradation. An alternative is to risk violating operational limits that lead to excessive fatigue or failure. Similarly, conventional power plant control systems lack the ability to respond to changing conditions most cost effectively. As will be appreciated, this results in power plant utilization that is often far from optimal. Therefore, there is a need for improved methods and systems for monitoring, modeling, and controlling the operation of power plants, among other things, the myriad operating modes available to operators of complex modern power plants, and their respective There is a need for methods and systems that allow a more complete understanding of the economic tradeoffs associated with operating modes.

U.S. Pat. No. 8816531

  Accordingly, the present application describes a control method for optimizing the operation of a power plant fleet. A power plant fleet may include remotely located assets for generating electricity for sale in the market. A power plant fleet can include multiple operating configurations that are differentiated by the manner in which the asset is involved, and the asset has multiple operating modes that can be described by operating parameters that relate to aspects of physical operation. are doing. The method comprises the steps of sensing and collecting measurements of operating parameters for each behavior of the asset, and tuning the asset model to form a tuned asset model for each of the assets. The steps include a data reconciliation adjustment process in which measured values for selected operating parameters are compared to predicted values for selected operating parameters to determine a difference between them. And the steps to tune the asset model are steps that are based on the difference and that use the tuned asset model to simulate the proposed operating configuration of the power plant fleet. , The simulation of the proposed motion configuration is The steps of defining a selected operating period, selecting a proposed operating configuration, and performing a simulation run for each of the proposed operating configurations using a tuned asset model, whereby , A step in which the operation of the power plant fleet during a selected operation period is simulated, including steps, and steps for obtaining simulation results from each of the steps and simulation runs, each of the simulation results being a performance index. , Including a predicted value for.

  These and other features of the present application will become more apparent when the following detailed description of the preferred embodiments is considered in conjunction with the drawings and the appended claims.

3 is a schematic diagram of a power system according to aspects of the invention. 3 is a schematic diagram of an exemplary thermal power generation unit as may be used in a power plant, according to embodiments of the invention. 1 is a schematic diagram of an exemplary power plant with multiple gas turbines, according to an embodiment of the invention. FIG. 3 illustrates an exemplary system configuration for a plant controller and optimizer, according to aspects of the present invention. 2 is a schematic diagram of a power plant with a plant controller and optimizer having a system configuration according to a particular aspect of the invention. FIG. 6 illustrates a computer system with an exemplary user interface according to certain aspects of the invention. FIG. 3 is a diagram showing an exemplary incremental heat rate curve and the effect that errors can have on an economical power supply process. 3 is a schematic diagram illustrating an exemplary plant controller with a power system in accordance with aspects of the present invention. 4 is a flow diagram of a power plant control method, according to an aspect of the invention. 3 is a dataflow diagram illustrating an architecture for a plant optimization system for a combined cycle power plant, according to aspects of the invention. FIG. 6 is a simplified block diagram of a computer system such as may be used with a real-time optimization system, according to aspects of the invention. 3 is a flow diagram of an exemplary method for solving parameterized simultaneous equations and constraints in accordance with the present invention. FIG. 6 is a diagram showing a simplified configuration of a computer system according to a control methodology of an embodiment of the present invention. FIG. 8 illustrates an alternative configuration of a computer system according to the control methodology of an embodiment of the present invention. 3 is a flow diagram of an exemplary control methodology, in accordance with an exemplary aspect of the present invention. 6 is a flow diagram of an alternative control methodology according to an exemplary aspect of the invention. 6 is a flow diagram of an alternative control methodology according to an exemplary aspect of the invention. FIG. 6 illustrates a flow diagram in which an alternative embodiment of the invention relating to turndown operation optimization is provided. FIG. 6 illustrates a flow diagram in which an alternative embodiment of the invention relating to optimizing between turndown and shutdown operations is provided. 4 is a diagram illustrating available operating modes of a gas turbine during selected operating periods having defined intervals, in accordance with aspects of an exemplary embodiment of the present invention. 6 is a diagram illustrating available operating modes of a gas turbine during selected operating periods having defined intervals in accordance with aspects of an alternative embodiment of the present invention. 6 is a flow diagram of a power plant fleet optimization process according to an alternative embodiment of the present invention. 2 is a schematic diagram of a power plant fleet optimization system according to aspects of the invention. 3 is a schematic diagram of a power plant fleet optimization system according to an alternative aspect of the invention. 3 is a schematic diagram of a power plant fleet optimization system according to an alternative aspect of the invention. 1 is a schematic diagram of a power block optimization system including a block controller. 7 is a schematic diagram of an alternative power block optimization system including a block controller. 3 is a flow chart illustrating an embodiment of a process for optimizing shutdown of a combined cycle power plant. FIG. 6 illustrates an exemplary control system in which a model-free adaptive controller is used in accordance with aspects of the present invention.

  Illustrative embodiments of the invention will now be described more fully hereinafter with reference to the accompanying drawings, in which some, but not all embodiments are shown. Indeed, the present invention can be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will satisfy applicable legal requirements. Like numbers may refer to like elements throughout.

  In accordance with aspects of the present invention, systems and methods are disclosed that can be used to optimize the performance of power systems, power plants, and / or thermal power generation units. In an exemplary embodiment, this optimization includes economic optimization, which allows the power plant operator to operate in either of the alternative modes of operation to enhance profitability. To decide. Embodiments may be utilized within a particular power system to provide a competitive advantage in obtaining an advantageous economic start-up plan during the power feed process. The advisor function may allow the operator to choose between operating modes based on accurate economic comparisons and expectations. As another feature, the process of prospectively purchasing fuel for future power generation periods may be improved such that fuel inventory is minimized without increasing the risk of starvation. Another aspect of the present invention provides a computer-implemented method and apparatus for modeling a power system and a power plant having a plurality of thermal power generation units, as described below. Technical effects of some configurations of the present invention include the generation and solution of energy system models that predict performance under various physical, operating, and / or economic conditions. Example embodiments of the present invention combine a power plant model that predicts performance under various ambient and operating conditions with an economic model that includes economic constraints, goals, and market conditions to improve profitability. It is designed to be optimized. In doing so, the optimization system of the present invention is characterized by ambient conditions, operating conditions, contractual conditions, regulatory conditions, legal conditions, and / or particular combinations of economic and market conditions. It is possible to predict an optimized set point that maximizes profitability.

  FIG. 1 illustrates a schematic diagram of a power system 10 including aspects of the invention, as well as an exemplary environment in which embodiments may operate. The power system 10 can include a generator or plant 12, such as, for example, the wind power and thermal power plants shown and the like. It will be appreciated that a thermal power plant may include a power generation unit such as a gas turbine, a coal fired steam turbine, and / or a combined cycle plant. In addition, the power system 10 may include other power sources such as solar power plants, hydropower sources, geothermal power sources, nuclear power sources, and / or any other suitable power source now known or to be discovered hereafter. It may include a power plant of the type (not shown). Power line 14 may connect various power plants 12 to customers or loads 16 of power system 10. It should be understood that the transmission line 14 represents a grid or distribution network for the power system and may include multiple sections and / or substations as desired or needed. is there. Electric power generated from the power plant 12 may be delivered to a load 16 via a transmission line 14, which may include, for example, a municipal customer, a residential customer, or a corporate customer. It is possible. The power system 10 can also include a storage device 18, which is connected to the power line 14 and is adapted to store energy during periods of over-generation.

  The power system 10 also includes a control system or controller 22, 23, 25, which controls or controls the operation of some of the components contained in the power system 10. . For example, the plant controller 22 can control each operation of the power generation plant 12. The load controller 23 is capable of controlling the operation of different loads 16 that are part of the power system 10. For example, the load controller 23 can manage the manner or timing of a customer's power purchase. The power dispatching authority 24 can manage certain aspects of the operation of the power system 10 and can also include a power system controller 25, which controls the economic powering procedure. A load start-up plan is distributed among the participating power plants according to the economic power supply procedure. The controllers 22, 23, 25 are represented by rectangular blocks, the controllers 22, 23, 25 can be connected to a communication network 20 via a communication line or a communication connection 21, by means of which the communication network 20 exchanges data. To be done. The connection 21 can be wired or wireless. It will be appreciated that the communication network 20 may be connected to, or part of, a larger communication system or network, such as the Internet or a private computer network. Additionally, the controllers 22, 23, 25 receive information, data, and instructions from and / or through the communication network 20 to and from the data library and data resources. Instructions may be sent and the data library and data resources may be generically referred to herein as "data resources 26", or alternatively, one or more such data resources. It is possible to store or store various data repositories locally. The data resources 26 can include several types of data, including, but not limited to, market data, operational data, and ambient data. Market data includes information about market conditions such as energy selling prices, fuel costs, labor costs, regulations, and so on. The operational data includes information related to operating conditions of the power plant or its power generation units, such as temperature or pressure measurements in the power plant, air flow rates, fuel flow rates, and the like. Ambient data includes information related to ambient conditions in the plant, such as ambient air temperature, humidity, and / or pressure. The market data, operational data, and ambient data may each include historical records, current condition data, and / or forecast-related data. For example, the data resource 26 may include current and forecast weather / climate information, current and forecast market conditions, historical usage and performance records for power plant operation, and / or similar components and / or configurations. It is possible to include parameters measured with respect to the operation of other power plants with, as well as other data as needed and / or desired. In operation, for example, the power system controller 25 of the power dispatching authority 24 receives data from other controllers 22, 23 in the power system 10 and issues instructions to the other controllers 22, 23 in the power system 10. It is possible to Each of the plant controller and load controller then controls the system components it is responsible for, relays information about the system components to the power system controller 25, and receives instructions from the power system controller 25.

  FIG. 2 is a schematic diagram of an exemplary thermal power generation unit or gas turbine system 30 that may be used in a power plant according to the present invention. As shown, the gas turbine system 30 includes a compressor 32, a combustor 34, a turbine 36 drivably coupled to the compressor 32, and a component controller 31. The component controller 31 can be connected to the plant controller 22, and the plant controller 22 can be connected to a user input device for receiving communication from the operator 39. It will be appreciated that, alternatively, the component controller 31 and the plant controller 22 can be combined into a single controller. The intake duct 40 guides ambient air to the compressor 32. As discussed in FIG. 3, injected water and / or other moisturizers may be directed to the compressor through intake duct 40. The air intake duct 40 may have filters, screens, and sound absorbing devices that contribute to the pressure loss of ambient air flowing through the air intake duct 40 and into the inlet guide vanes 41 of the compressor 32. The exhaust duct 42 guides the combustion gas from the outlet of the turbine 36, for example through an emission control and sound absorbing device. The sound absorbing material and the emission control device can apply back pressure to the turbine 36. The turbine 36 may drive a generator 44 that produces electrical power, which may then be distributed through the electrical power system 10 via the transmission line 14.

  The operation of the gas turbine system 30 can be monitored by a number of sensors 46, which detect various operating conditions or operating parameters throughout the gas turbine system 30, which, for example, , Compressor 32, combustor 34, turbine 36, generator 44, and conditions in ambient environment 33. For example, the temperature sensor 46 can monitor ambient temperature, compressor discharge temperature, turbine exhaust temperature, and other temperatures in the flow path of the gas turbine system 30. Similarly, pressure sensor 46 can monitor static pressure and dynamic pressure levels at ambient pressure, compressor inlet, compressor outlet, turbine exhaust, and other suitable locations within the gas turbine system. . Humidity sensor 46, such as a wet bulb thermometer and a dry bulb thermometer, is capable of measuring ambient humidity in the compressor intake duct. Also, the sensor 46 is typically used to measure flow sensors, velocity sensors, flame detector sensors, valve position sensors, guide vane angle sensors, and various operating parameters and operating conditions for the operation of the gas turbine system 30. It is possible to include other sensors used. As used herein, the term "parameter" is used to define operating conditions within a system, such as gas turbine system 30 or other power generation system described herein. Represent measurable physical characteristics of motion that may be used for. Operating parameters include temperature, pressure, humidity, and gas flow characteristics at locations defined along the working fluid path, as well as ambient conditions, fuel characteristics, and other measurable as may be suitable without limitation. Properties can be included. It will also be appreciated that the control system 31 includes a number of actuators 47 by which the control system 31 mechanically controls the operation of the gas turbine system 30. The actuator 47 is a variable variable actuator that allows manipulation of a particular process input (ie, manipulated variable) for control of a process output (ie, controlled variable) according to a desired outcome or mode of operation. It is possible to include electromechanical devices with set points or set points. For example, the commands generated by the component controller 31 may be fuel supplies where one or more actuators 47 in the turbine system 30 regulate flow levels, fuel splits, and / or the type of fuel burned. It is possible to trigger adjusting the valve between the and combustor 34. As another example, the command generated by the control system 31 can cause one or more actuators to adjust inlet guide vane settings that change the orientation angle of the inlet guide vanes.

  The component controller 31 can be a computer system having a processor that uses sensor measurements and instructions from a user or plant operator (hereinafter "operator 39") to gas turbine. The program code is executed to control the operation of system 30. As discussed in more detail below, the software executed by controller 31 may include a scheduling algorithm for coordinating any of the subsystems described herein. The component controller 31 is capable of adjusting the gas turbine system 30 based in part on algorithms stored in its digital memory. These algorithms allow, for example, the component controller 31 to maintain NOx and CO emissions in turbine exhaust within certain predetermined emission limits, or in other cases, combustor combustion. It may allow the temperature to be maintained within predetermined limits. It is recognized that the algorithm can include inputs for parameter variables such as compressor pressure ratio, ambient humidity, inlet pressure drop, turbine exhaust back pressure, etc., as well as any other suitable parameter. Will be done. The schedule and algorithms executed by the component controller 31 adapt to changing ambient conditions, which affect emissions, combustor dynamics, and combustion temperature limits, such as at full load and partial load operating conditions. Exert. As discussed in more detail below, the component controller 31 sets the desired turbine exhaust temperature and combustor fuel split with the goal of meeting performance goals while adhering to the operability limits of the gas turbine system. It is possible to apply algorithms for scheduling gas turbines, such as those that do. For example, the component controller 31 can determine the combustor temperature rise and NOx during partial load operation, increasing the operating margin to the combustion dynamics limit, thereby improving the operability and reliability of the power generation unit. , And to improve operability.

  Turning to FIG. 3, a schematic diagram of an exemplary power plant 12 having a plurality of power generation units or plant components 49 in accordance with aspects of the present invention is provided. The illustrated power plant 12 of FIG. 3 is a common configuration and will therefore be used to discuss some of the exemplary embodiments of the invention presented below. However, it will be appreciated that the methods and systems described herein are more widely applicable and applicable to power plants having more power generation units than the power generation units shown in FIG. It may be expandable, while still applicable to power plants with a single power generation component such as the one illustrated in FIG. It will be appreciated that the power plant 12 of FIG. 3 is a combined cycle plant, which includes a number of plant components 49, including a gas turbine system 30 and a steam turbine system 50. Power generation may be augmented by other plant components 49, such as an intake air conditioning system 51 and / or a heat recovery steam generator with a duct firing system (hereinafter "HRSG duct firing system 52"). . Each of the gas turbine system 30, the steam turbine system 50 including the HRSG duct firing system 52, and the intake conditioning system 51 is a control system or component controller that electronically communicates with sensors 46 and actuators 47 dedicated to the respective plant components. It will be recognized that 31 is included. As used herein, the intake conditioning system 51 can represent components used to condition air prior to entering the compressor, unless otherwise stated. , An intake chilling system or chiller, an evaporator, an atomizer, a water injection system, and / or in some alternative cases, a heating element.

  In operation, the intake conditioning system 51 is adapted to cool the air entering the gas turbine system 30 and enhance the power generating capacity of the unit. The HRSG duct firing system 52 is adapted to burn fuel, provide additional heat, and increase the supply of steam expanded through the turbine 53. As such, the HRSG duct firing system 52 enhances the energy provided by the hot exhaust gas 55 from the gas turbine system, thereby increasing the power generation capacity of the steam turbine system.

  As an exemplary operation, power plant 12 of FIG. 3 directs a flow of fuel to combustor 34 of gas turbine system 30 for combustion. The turbine 36 is powered by combustion gases and drives a compressor 32 and a generator 44, which delivers electrical energy to the transmission line 14 of the power system 10. The component controller 31 of the gas turbine system 30 sets the commands for the gas turbine system regarding fuel flow rate, and gas turbine system, such as air inlet temperature, humidity, power output, shaft speed, and exhaust gas temperature. It is possible to receive sensor data from the system. The component controller 31 can also collect other operational data from pressure and temperature sensors, flow control devices, and other devices that monitor the operation of the gas turbine system. The component controller 31 can send data regarding the operation of the gas turbine system and receive instructions from the plant controller 22 regarding set points for actuators that control the process inputs.

  Air entering the gas turbine system 30 during a particular mode of operation may be cooled by the intake conditioning system 51 or otherwise conditioned to enhance the power generation capacity of the gas turbine system. . The intake conditioning system 51 may include a refrigeration system 65 for cooling water and a component controller 31 that controls the operation of the intake conditioning system 51. In this case, the component controller 31 can receive information regarding the temperature of the cooling water, as well as instructions regarding the level of injection desired, which instructions can come from the plant controller 22. The component controller 31 of the intake conditioning system 51 can also issue commands that cause the refrigeration system 65 to produce cooling water having a particular temperature and flow rate. The component controller 31 of the intake conditioning system 51 can send data regarding the operation of the intake conditioning system 51.

  The steam turbine system 50 may include a turbine 53 and a HRSG duct firing system 52, and a component controller 31 dedicated to controlling the operation of the steam turbine system 50, as shown. Hot exhaust gases 55 from the exhaust ducts of gas turbine system 30 can be directed into steam turbine system 50 to create steam, which is expanded through turbine 53. As will be appreciated, the HRSG duct firing system is constantly used to provide additional energy for the production of steam to increase the power generation capacity of the steam turbine system. The rotation induced in the turbine 53 by the steam is adapted to drive the generator 44 and produce electrical energy, which is then sold in the power system 10 across the transmission line 14. Will be recognized. The component controller 31 of the steam turbine system 50 can set the flow rate of fuel burned by the duct firing device 52, thereby increasing steam production beyond what can be produced by the exhaust gas 55 alone. Is. The component controller 31 of the steam turbine system 50 can send data regarding the operation of the plant components 49 and receive instructions from that data regarding how the steam turbine system 50 should operate.

  The plant controller 22 of FIG. 3 is connected to each of the component controllers 31, as shown, and is capable of communicating with the sensors 46 and actuators 47 of several plant components 49 via these connections. is there. As part of controlling the power plant 12, the plant controller 22 can simulate the operation of the power plant 12. More specifically, the plant controller 22 may include or be in communication with a digital model (or simply “model”), which model simulates the operation of each plant component 49. To do. The model can include an algorithm that correlates process input variables with process output variables. An algorithm can include a set of instructions, logic, mathematical formulas, functional relationship descriptions, schedules, data collection, and so on. In this case, the plant controller 22 includes a gas turbine model 60 that models the operation of the gas turbine system 30, an intake conditioning system model 61 that models the operation of the intake conditioning system 51, the steam turbine system 50 and the HRSG duct. A steam turbine model 62 that models the operation of the firing system 52. As a general note, the system and its associated model, as well as the discrete steps of the methods provided herein, may be subdivided in various ways without substantially departing from the scope of the invention, It will be appreciated that and / or that they may be combined and the manner in which each is described is exemplary, unless otherwise stated or claimed. Using these models, the plant controller 22 can simulate the operation of the power plant 12, such as parameters describing thermodynamic performance or operation.

  The plant controller 22 can then use the results from the simulation and are adapted to determine the optimized operating mode. Such optimized modes of operation may be described by a parameter set, which includes a plurality of operating parameters and / or set points for actuators, and / or other operating conditions. As used herein, an optimized mode of operation is at least a preferred mode of operation over at least one alternative mode of operation according to a defined standard or performance index, and the defined standard of operation. Alternatively, performance indicators may be selected by the operator to evaluate plant operation. More specifically, the optimized mode of operation is preferred over one or more other possible modes of operation as used herein and simulated by the plant model. Is an operation mode evaluated as. The optimized mode of operation is determined by assessing how the model predicts that the power plant will operate under each mode of operation. As discussed below, an optimizer 64, eg, a digital software optimization program, was run to run the digital power plant model according to various parameter sets, and then preferred or optimized by evaluating the results. It is possible to specify the mode of operation. The change in set point can be caused by a perturbation applied around the set point chosen for the analysis. These can be based in part on historical activity. It will be appreciated that the optimized mode of operation may be determined by optimizer 64 based on one or more defined cost functions. Such a cost function may consider, for example, the cost of producing electricity, profitability, efficiency, or some other criterion as defined by operator 39.

  To determine cost and profitability, the plant controller 22 can include or be in communication with an economic model 63, which can be, for example, a gas turbine system, an intake conditioning system, And the price of electricity and certain other variable costs, such as the cost of fuel used in the HRSG duct firing system. The economic model 63 provides the data used by the plant controller 22 to propose proposed set points (ie, the selected set points, the behavior of which is to determine the optimized set points). It is possible to determine which of the (modeled) represents the minimum production cost or maximum profitability. According to other embodiments, as discussed in more detail in FIG. 4, the optimizer 64 of the plant controller 22 includes a filter, such as a Kalman filter, or works with such a filter. It can assist in tuning, adjusting, and calibrating the digital model such that the model accurately simulates the operation of the power plant 12. As discussed below, the model can be a dynamic model, which includes a learning mode in which the actual operation (ie, the actual operation of the power plant 12 is The model is tuned or reconciled through a comparison made between the measured behavioral parameter (which reflects the value of) and the predicted behavior (ie, the value for the same behavioral parameter that the model predicted). (Reconciled). Also, as part of the control system, the filter can be used to adjust or calibrate the model in real time, or near real time, eg, every few minutes or hours, or as specified.

  The optimized set points generated by the plant controller 22 represent recommended modes of operation, such as fuel and air settings for gas turbine systems, temperature and water mass flow rates for intake conditioning systems, A level of duct firing within the steam turbine system 50 may be included. According to particular embodiments, these suggested operating set points may be provided to operator 39 via an interface device, such as a computer display screen, printer, or sound speaker. By knowing the optimized set point, the operator can then input the set point into the plant controller 22 and / or the component controller 31, which plant controller 22 and / or the component controller 31 Then, control information for realizing the recommended mode of operation is generated. In embodiments where the optimized set point does not include specific control information to achieve the operating mode, the component controller can provide the control information needed for this, and more As discussed below, the component controller can continue to control plant components in a closed loop fashion according to the recommended mode of operation until the next optimization cycle. Also, depending on the operator's choice, the plant controller 22 can implement the optimized setpoints directly or automatically without operator involvement.

  As an exemplary operation, power plant 12 of FIG. 3 directs a flow of fuel to combustor 34 of gas turbine system 30 for combustion. The turbine 36 is powered by combustion gases and drives a compressor 32 and a generator 44, which delivers electrical energy to the transmission line 14 of the power system 10. The component controller 31 sets commands for the gas turbine system 30 regarding fuel flow rate and from the gas turbine system 30 such as air inlet temperature and humidity, power output, shaft speed, and exhaust gas temperature. It is possible to receive sensor data. The component controller 31 can also collect other operational data from pressure and temperature sensors, flow control devices, and other devices that monitor the gas turbine system 30. The component controller 31 of the gas turbine system 30 is capable of sending data regarding the operation of the system and receiving set point instructions for the actuators controlling the process inputs from the plant controller 22.

  During certain modes of operation, the air entering the gas turbine system 30 may be cooled by chilled water supplied from the intake conditioning system 51 to the inlet air duct 40. It will be appreciated that cooling the air entering the gas turbine can be done to enhance the gas turbine engine's ability to generate electrical power. The intake air conditioning system 51 can include a refrigeration system or refrigerator 65 for cooling water and a component controller 31. In this case, the component controller 31 receives information regarding the temperature of the cooling water and a command regarding the desired cooling of the intake air. These commands can come from the plant controller 22. The component controller 31 of the intake conditioning system 51 can also issue commands that cause the refrigeration system 65 to produce cooling water having a particular temperature and flow rate. The component controller 31 of the intake conditioning system 51 can send data regarding the operation of the intake conditioning system 51 and receive instructions from the controller 22.

  The steam turbine system 50 may include an HRSG 52 with a duct firing device, a steam turbine 53, and a component controller 31 that may be dedicated to the operation of the steam turbine system 50. The hot exhaust gas 55 from the exhaust duct 42 of the gas turbine system 30 is directed into the steam turbine system 50 and produces steam that drives the steam turbine system 50. The HRSG duct firing system 52 is used to provide additional heat energy, capable of producing steam, and adapted to increase the power generation capacity of the steam turbine system 50. The steam turbine 53 drives the generator 44 to produce electrical energy, which is delivered to the power system 10 via the power transmission line 14. The component controller 31 of the steam turbine system 50 can set the flow rate of the fuel burned by the duct firing device 52. The heat generated by the duct firing device exceeds the amount created by the exhaust gases 55 from the turbine 36 alone, increasing steam generation. The component controller 31 of the steam turbine system 50 can send data regarding the operation of the system and receive instructions from the plant controller 22.

  The plant controller 22 can communicate with an operator 39 and data resources 26 to receive data about market conditions, such as price and demand for delivered power. According to certain embodiments, plant controller 22 issues recommendations to operator 39 regarding desired operating set points for gas turbine system 30, intake conditioning system 51, and steam turbine system 50. The plant controller 22 is capable of receiving and storing data about the operation of the power plant 12 components and subsystems. The plant controller 22 can be a computer system having a processor and memory that stores data, digital models 60, 61, 62, 63, optimizer 64, and other computer programs. The computer system may be embodied in a single physical or virtual computing device, or distributed across local or remote computing devices. The digital models 60, 61, 62, 63 may be embodied as a set of algorithms, such as transfer functions, that relate to respective operating parameters of the system. The model can include a physically based aerodynamic computer model, a regression fit model, or other suitable computer-implemented model. According to a preferred embodiment, the models 60, 61, 62, 63 are tuned, adjusted or calibrated or predicted on a regular, automatic and real-time or near real-time basis. Can be tuned according to an ongoing comparison between the measured and actual measured parameters. The models 60, 61, 62, 63 can include a filter, which receives data input regarding the actual physical and thermodynamic operating conditions of the combined cycle power plant. These data inputs may be provided to the filter during operation of the power plant 12, in real time or periodically, every 5 minutes, every 15 minutes, every hour, every day, etc. The data input can be compared to the data predicted by the digital models 60, 61, 62, 63, and based on the comparison, the model can be continuously refined.

  FIG. 4 illustrates a schematic system configuration of the plant controller 22, which includes a filter 70, an artificial neural network configuration 71 (“neural network 71”), and an optimizer 64 according to aspects of the invention. . The filter 70 can be, for example, a Kalman filter, and the filter 70 simulates the operation of the power plant 12 with actual data 72 of measured operating parameters from the sensor 46 of the power plant 12. It is possible to compare the data 60 with the same operating parameters predicted by the models 60, 61, 62, 63 and the neural network 71. The difference between the actual and predicted data can then be used by the filter 70 to tune the model of the power plant simulated by the neural network 71 and the digital model.

  Although certain aspects of the present invention are described herein with reference to models in the form of neural network based models, the present invention is not limited thereto, but includes physics-based models, data driven models. Others, including models, empirically developed models, heuristics-based models, support vector machine models, models developed by linear regression, models developed using knowledge of "first principles", etc. It should be understood that it is expected that it can be implemented using a model of Additionally, in order to properly capture the relationship between manipulated / disturbed variables and controlled variables, according to certain preferred embodiments, the power plant model has one of the following characteristics: It is possible to have one or more. 1) Non-linearity (a non-linear model can represent a curve rather than a straight line relationship between manipulated / disturbed variables and controlled variables); 2) multiple inputs / multiple Outputs (models can capture relationships between multiple inputs (manipulated and disturbance variables) and multiple outputs (controlled variables); 3) Being dynamic (of inputs It is possible that a change does not momentarily affect the output, but rather there is a time delay, and a time delay is followed by a dynamic response to the change, for example a change in input across the system. It can take a few minutes for it to propagate fully, and since the optimization system runs at a given frequency, the model represents the effect of these changes over time and must be taken into account. 4) Suitable That the model is updated (the model can be updated at the start of each optimization to reflect current operating conditions); and 5) derived from empirical data (each power plant Is unique, the model can be derived from empirical data obtained from the power generation unit). Given the above requirements, a neural network based approach is the preferred technique for implementing the required plant model. Neural networks can be developed based on empirical data using advanced regression algorithms. As will be appreciated, neural networks can capture the non-linearities commonly shown in the operation of power plant components. Also, neural networks can be used to represent systems with multiple inputs and outputs. In addition, neural networks can be updated using either feedback bias or online adaptive learning. Also, the dynamic model may be implemented in a neural network based structure. A variety of different types of model architectures have been used for dynamic neural network implementations. Many neural network model architectures require large amounts of data to successfully train dynamic neural networks. Assuming a robust power plant model, it is possible to calculate the effect of changes in manipulated variables on the controlled variables. Moreover, because the plant model is dynamic, it is possible to calculate the effect of changes in the manipulated variables over future planning horizon.

  The filter 70 can generate a performance multiplier applied to the inputs or outputs of the digital model and the neural network, or the weighting applied to the logic units and the algorithm used by the digital model and the neural network. Can be modified. These effects of the filter reduce the discrepancy between the actual and predicted data. The filter continues to operate to further reduce the discrepancy or to address possible variations. By way of example, filter 70 may include compressor discharge pressure and temperature in a gas turbine, gas turbine and steam turbine efficiency, gas turbine systems, intake conditioning systems, and fuel flow to HRSG duct firing systems, and / or , It is possible to generate performance multipliers for the predicted data on other suitable parameters. It will be appreciated that these categories of operating data reflect operating parameters that experience performance degradation over time. By providing performance multipliers for these types of data, filter 70 may be particularly useful in adjusting models and neural networks to account for performance degradation of power plants.

  As shown in FIG. 4, according to a particular embodiment of the invention, each of the digital models 60, 61, 62, 63 of some plant components 49 of the power plant of FIG. , The algorithm is represented by several graphs, which are used to model the corresponding systems. It will be appreciated that the model interacts and communicates with the neural network 71, in which the neural network 71 forms the overall model of the combined cycle power plant 12. In this way, the neural network simulates the thermodynamic and economic behavior of the plant. As indicated by the solid arrows in FIG. 4, the neural network 71 collects the data output by the models 60, 61, 62, 63, and the data that will be used as input by the digital model. I will provide a.

  The plant controller 22 of FIG. 4 also includes an optimizer 64, such as a computer program, which interacts with the neural network 71 to provide a gas turbine system, an intake conditioning system, a steam turbine system, and an HRSG duct. Find the optimal set point for the firing system and achieve the defined performance goals. The performance goal can be, for example, maximizing the profitability of the power plant. The optimizer 64 can cause the neural network 71 to run digital models 60, 61, 62, 63 at various operating set points. The optimizer 64 can have a perturbation algorithm, which assists in changing the operating set point of the model. The perturbation algorithm causes the simulation of the combined cycle power plant provided by the digital model and the neural network to operate at a set point that is different from the current operating set point for the plant. By simulating the operation of the power plant at different set points, the optimizer 64 causes the operation set points that will cause the plant to operate more economically, or some other that may be defined by the operator 39. Search for an action set point that will improve performance according to the criterion.

  According to an exemplary embodiment, economic model 63 provides data used by optimizer 64 to determine which setpoint is the most profitable. The economic model 63 may receive and store formatted fuel cost data, such as a chart 630 that correlates fuel costs over time, such as during each season of the year. Another chart 631 can correlate the price received for electricity at different times of the day, week, or month. The economic model 63 can provide data regarding the price received for electricity and the cost of fuel used to produce electricity (gas turbine fuel, duct firing fuel, and intake conditioning system fuel). Is. Data from the economic model 63 can be used by the optimizer 64 to evaluate each of the operating conditions of the power plant according to the performance goals defined by the operator. The optimizer 64 determines which of the operating conditions of the power plant 12 is optimal given the performance goals as defined by the operator 39 (which, as used herein, may be It means that it is at least preferable to the operating state). The digital model, as described, is used to simulate the operation of plant components 49 of the power plant 12, such as modeling the thermodynamic operation of a gas turbine system, an intake conditioning system, or a steam turbine system. Can be done. The model can include algorithms, such as mathematical equations and lookup tables, which are stored locally, updated periodically, or obtained remotely via data resource 26. In effect, it simulates the response of the plant component 49 to specific input conditions. Such a look-up table can include measured operating parameters that describe the same type of operation of components operating in a remote power plant installation.

  The thermal model 60 of the gas turbine system 30 includes, for example, an algorithm 600 that correlates the effects of inlet air temperature with power output. It will be appreciated that this algorithm can indicate that the power output decreases from the maximum value 601 as the inlet air temperature increases above the threshold 602 temperature. The model 60 can also include an algorithm 603 that correlates gas turbine heat rates at different engine power output levels. As discussed, heat rate represents the efficiency of a gas turbine engine or other power generation unit and is inversely proportional to efficiency. Lower heat rates indicate higher thermodynamic performance efficiencies. The digital model 61 is capable of simulating the thermodynamic behavior of the intake conditioning system 51. In this case, for example, the digital model 61 includes an algorithm 610, which correlates the chilling capacity based on the energy applied to run the refrigeration system 65 of the intake conditioning system 51, and calculates the chilling capacity. Is to indicate the amount of cooling applied to the air entering the gas turbine. There may be a maximum chilling capacity value 611 that may be achieved by refrigeration system 65. In another case, the associated algorithm 612 can correlate the energy applied to run the refrigeration system 65 with the temperature of the cold air entering the compressor 32 of the gas turbine system 30. The model 61 shows that, for example, reducing the temperature of the air entering the gas turbine below the dew point 613 of the ambient air dramatically increases the power required to run the intake conditioning system. It is possible to show. In the case of the steam turbine system 50, the digital model 62 may include an algorithm 620, such as the power output of the steam turbine system, such as the amount of fuel consumed by duct firing, and the like. Correlates to the energy added by the HRSG duct firing system 52. The model 62 can show that there can be an upper threshold level 621 for an increase in steam turbine system power output that can be achieved, for example, by an HRSG duct firing system, which is in the algorithm 620. May be included.

  According to a particular embodiment of the invention, as illustrated in FIG. 4, a neural network 71 interacts with each of the digital models of several plant components 49 of the power plant 12 of FIG. It is possible to provide communication between. The interaction can include collecting output data from the model and generating input data that is used by the model to generate additional output data. The neural network 71 can be a digital network connected to logic elements. The logic elements may each implement an algorithm that accepts a data input and produces one or more data outputs. Simple logic elements can sum the values at the inputs to produce output data. Other logic elements can multiply the value at the input or apply other mathematical relationships to the input data. The data input to each of the logic elements of neural network 71 may be assigned a weighting such as a multiplier between 1 and zero. The weighting can be modified during the learning mode, which adjusts the neural network to better model the performance of the power plant. Also, the weighting may be adjusted based on the commands provided by the filter. Adjusting the weighting of the data inputs to the logic units in the neural network is one example of how the neural network may be dynamically modified during operation of a combined cycle power plant. Another example is to modify the weighting of data inputs to algorithms (which are examples of logic units) in each of thermodynamic digital models for steam turbine systems, intake conditioning systems, and gas turbines. including. The plant controller 22 may be modified in other ways, such as adjustments made to the logic units and algorithms based on the data provided by the optimizer and / or filters.

  The plant controller 22 can generate an output of a recommended or optimized set point 74 for the combined cycle power plant 12, which output, as shown, is a power plant actuator 47. It is possible to pass through the operator 39 for approval before being transmitted to and implemented by the power plant actuator 47. As shown, the optimized set point 74 includes input from an operator 39 via a computer system, such as the one described below in connection with FIG. 6, or the operator Approved by 39. The optimized set point 74 is, for example, the temperature and mass flow rate associated with the cooling water generated by the intake conditioning system and used to cool the air entering the gas turbine system; the fuel flow rate to the gas turbine system. As well as a duct firing rate. Also, the optimized set point 74 can then be used by the neural network 71 and the models 60, 61, 62, 63 so that ongoing plant simulations can predict operational data. It will be appreciated that the data may later be compared to actual operating data, allowing the plant model to be continuously refined.

  FIG. 5 illustrates a simplified system configuration of the plant controller 22 with the optimizer 64 and power plant model 75. In this exemplary embodiment, plant controller 22 includes optimizer 64 and power plant model 75 (which may be, for example, neural network 71 and models 60, 61, 62, 63 discussed above in connection with FIG. 4). Are included). The power plant model 75 can simulate the overall operation of the power plant 12. According to the illustrated embodiment, the power plant 12 includes a plurality of power generation units or plant components 49. The plant components 49 can include, for example, thermal power generation units, or other plant subsystems as previously described, each of which can include a corresponding component controller 31. . The plant controller 22 is capable of communicating with the component controller 31 and controls the operation of the power plant 12 through the component controller 31, and by the component controller 31 via connections to the sensors 46 and actuators 47. Is possible.

  It will be appreciated that a power plant has a number of variables that affect its operation. Each of these variables can generally be classified as either an input variable or an output variable. Input variables represent process inputs and include variables that can be manipulated by the plant operator, such as air and fuel flow rates. Input variables also include variables that cannot be manipulated, such as ambient conditions. Output variables are variables, such as power output, that are controlled by manipulating input variables that can be manipulated. The power plant model is configured to represent algorithmic relationships between input variables and output variables or controlled variables, which can be manipulated variables or "manipulated variables" and non-manipulated variables. Alternatively, an output variable or a controlled variable including a “disturbance variable” will be referred to as a “controlled variable”. More specifically, the manipulated variable is a variable that can be changed by the plant controller 22 to affect the controlled variable. Variables that are manipulated include things like valve set points that control fuel and air flow. Disturbance variables represent variables that affect the controlled variable but cannot be manipulated or controlled. Disturbance variables include ambient conditions, fuel characteristics, and the like. The optimizer 64 (1) meets performance goals of the power plant (eg, meets load requirements while maximizing profitability), and (2) constraints associated with operating the power plant (eg, emissions). And equipment restrictions), determine an optimal set of setpoint values for the manipulated variables.

  According to the present invention, the "optimization cycle" can be started at a predetermined frequency (for example, every 5 to 60 seconds, or every 1 to 30 minutes). At the start of the optimization cycle, the plant controller 22 sends current data on manipulated, controlled and disturbance variables directly from the component controller 31 and / or directly from the respective sensor 46 of the plant component 49. It is possible to obtain The plant controller 22 can then use the power plant model 75 to determine optimal setpoint values for the manipulated variables based on current data. In doing so, the plant controller 22 can run the plant model 75 at various operating set points, and which set of operating set points is most preferable given the performance goals of the power plant. , Which can be referred to as a “simulation run”. For example, the performance goal may be to maximize profitability. By simulating the operation of the power plant at different set points, the optimizer 64 allows the plant model 75 to predict that the plant model 75 will cause the plant to operate optimally (or at least in a preferred manner). Explore the set. As stated, this optimal set of set points may be referred to as an "optimized set point" or "optimized operating mode." Typically, in reaching the optimized setpoint, the optimizer 64 will compare multiple sets of setpoints, which will be subject to operator-defined performance goals. Then it will be found to be superior to each of the other sets. The operator 39 of the power plant 12 may have the option to approve the optimized set point, or the optimized set point may be approved automatically. The plant controller 22 can send the optimized set point to the component controller 31 or, alternatively, directly to the actuator 47 of the plant component 49, where the set point is the optimized setting. It can be adjusted according to the points. The plant controller 22 can be run in closed loop and is operated at a predetermined frequency (eg, every 10 to 30 seconds, or more frequently) based on the measured current operating conditions. It adjusts the value of the set point of the variable.

  The optimizer 64 can be used to minimize the "cost function" that will be subject to the set of constraints. The cost function is essentially a mathematical representation of plant performance goals, and the constraints are the limits within which the power plant must operate. Such limits can represent legal, regulatory, environmental, equipment, or physical constraints. For example, to minimize NOx, the cost function includes terms that decrease as the level of NOx decreases. One common method for minimizing such cost functions is known as, for example, "gradient descent optimization". Gradient descent is an optimization algorithm that approaches the local minimum of a function by taking a step proportional to the negative number of the gradient (or approximate gradient) of the function at the current point. It should be appreciated that multiple different optimization techniques may be used depending on the model morphology, as well as costs and constraints. For example, it is anticipated that the present invention may be implemented using various different types of optimization approaches, either individually or in combination. These optimization approaches include, but are not limited to, linear programming, quadratic programming, mixed integer nonlinear programming, stochastic programming, global nonlinear programming, genetic algorithms, and particle swarm optimization techniques. Including. Additionally, the plant model 75 can be dynamic so that the effects of changes are taken into account over future planned horizon. Therefore, the cost function includes terms over the future planned horizon. This approach is referred to as model predictive control because the model is used to make predictions over the time horizon for planning, which is described by S. Piche, B. Sayyar-Rodsari, D. Johnson and M. Gerules, "Nonlinear model predictive. control using neural networks, "IEEE Control Systems Magazine, vol. 20, no. 2, pp. 53-62, 2000, which is fully incorporated herein by reference.

  Constraints may be placed on both the process input (which includes manipulated variables) and the process output (which includes controlled variables) of the power plant over future planned periods. Typically, constraints are placed on the manipulated variables that match the limits associated with the plant controller. Output constraints can be determined by the problem being solved. According to an embodiment of the invention, and as a step in the optimization cycle, the optimizer 64 calculates a complete trajectory of the movement of the manipulated variable over a future planned period of time, eg 1 hour. It is possible. Therefore, for an optimized system that runs every 30 seconds, 120 values may be calculated for each manipulated variable over a one-hour future planned period. Since the plant model, or performance goals, or constraints may change before the next optimization cycle, the plant controller 22 / optimizer 64 will set the optimized set point for each manipulated variable. It is possible to output only the first value in the planned period for each manipulated variable to the component controller 31. In the next optimization cycle, the plant model 75 may be updated based on the current conditions. Also, the cost function and constraints can be updated if they changed. The optimizer 64 may then be used to recalculate the set of values for the manipulated variables over the time horizon, with the first value in the time horizon for each variable manipulated. It is output to the component controller 31 as the set point value for the manipulated variable. The optimizer 64 can repeat this process for each optimization cycle, which causes the power plant 12 to be affected by unexpected changes in such items as load, ambient conditions, fuel characteristics, and the like. Maintain optimal performance at a constant rate.

  Turning to FIG. 6, an exemplary environment and user input device for a plant controller and control program is illustrated in accordance with an exemplary embodiment. Embodiments include a computer system 80 having a display 81, a processor 82, a user input device 83, and a memory 84, although other configurations are possible. Aspects of computer system 80 may be located at power plant 12, while other aspects may be remote and connected via communication network 20. As discussed, the computer system 80 may be connected to each power generation unit or other plant component 49 of the power plant 12. Power plant component 49 may include gas turbine system 30, steam turbine system 50, intake conditioning system 51, HRSG duct firing system 52, and / or any subsystems or subcomponents associated therewith, or any of them. Combinations can be included. Also, computer system 80 may be connected to one or more sensors 46 and actuators 47 as needed or desired. As mentioned, the sensor 46 may be configured to sense the operating conditions and parameters of the components and relay signals to the computer system 80 regarding these conditions. Computer system 80 may be configured to receive these signals and use them in the manner described herein, which is to send signals to one or more of actuators 47. Can be included. However, unless otherwise required, the present invention may include embodiments that are not configured to directly control the power plant 12 and / or sense operating conditions. In the inventive arrangements for controlling power plant 12 and / or sensing operating conditions, such inputs or controls interact more directly with the physical components of the power plant and its sensors and actuators. It may be provided by receiving and / or transmitting signals to / from one or more separate software or hardware systems. Computer system 80 can include a power plant control program (“control program”), which manages data in a plant controller by performing the processes described herein. Computer system 80 as described above.

  Generally, processor 82 executes program code defining a control program, which is at least partially fixed in memory 84. While executing the program code, the processor 82 may process the data, which may include reading the converted data from the memory 84 and / or writing the converted data to the memory 84. It can result. The display 81 and input device 83 allow a human user to interact with the computer system 80 and / or one or more communication devices, allowing the system user to interact with the computer system 80 using any type of communication link. It may be possible to communicate. In an embodiment, a communication network, such as networking hardware / software, etc., may allow computer system 80 to communicate with other devices, both internal and external to the node in which the communication network is installed. To this extent, the control program of the present invention is capable of managing a set of interfaces that allow a human user and / or a system user to interact with the control program. Further, the control program manages data such as control data (eg, storing, reading, generating, manipulating, organizing, using any solution, as discussed below). It is possible to present).

  Computer system 80 may include one or more general purpose computing products capable of executing program code, such as control programs as defined herein, the program code being computer system 80. Installed on top of. As used herein, "program code" means any collection of instructions written in any language, code, or notation, where an instruction is a computing capable of processing information. The device is directly or after any combination of (a) conversion to another language, code or notation, (b) reproduction in different material forms, and / or (c) reconstruction. , It is understood to cause carrying out a particular action. Additionally, computer code can include object code, source code, and / or executable code, and also forms part of a computer program product when on at least one computer readable medium. It is possible to The term "computer-readable medium" may include one or more of any type of tangible medium of expression now known or later developed, in which a copy of the program code is tangible. It is understood that from the expression medium, it may be perceived, reproduced, or otherwise communicated by a computing device. When the computer executes the computer program code, the computer becomes a device for carrying out the present invention, and on a general-purpose microprocessor, specific logic circuits are generated by the configuration of the microprocessor including the computer code segment. . The technical effect of the executable instructions is to implement the power plant control method and / or system and / or computer program product, which is based on assumed ambient and / or market conditions, performance parameters, and / or Given the life cycle costs associated with it, models are being used to enhance, enhance, or optimize the operating characteristics of a power plant to more efficiently utilize the economic benefits of the power plant. . In addition to using current information, it is possible to use historical and / or predictive information, and a feedback loop allows the plant to operate more efficiently and dynamically during changing conditions. Can be established as The computer code of the control program may be written in computer instructions executable by the plant controller 22. To this extent, the control program executed by computer system 80 may be embodied as any combination of system software and / or application software. Further, the control program may be implemented using a set of modules. In this case, the modules may enable computer system 80 to perform the set of tasks used by the control program, and the modules may be separately developed and / or of the control program. It can be implemented separately from other parts. As used herein, the term “component” means any configuration of hardware, with or without software, that is used in conjunction with any component using any solution. The term "module" means program code that enables a computer system, using any solution, to implement the acts described in conjunction with the modules. are doing. When fixed in memory 84 of computer system 80, including processor 82, modules are substantive parts of the components that implement the operations. Nevertheless, it will be appreciated that two or more components, modules, and / or systems may share some / all of their respective hardware and / or software. Further, it is understood that some of the functionality discussed herein may not be implemented or additional functionality may be included as part of computer system 80. . When computer system 80 includes multiple computing devices, each computing device may have only a portion (eg, one or more modules) of a control program fixed on the computing device. Is. Regardless, when computer system 80 includes multiple computing devices, the computing devices can communicate by any type of communication link. Moreover, computer system 80 can communicate with one or more other computer systems using any type of communication link while performing the processes described herein. .

  The control program, as discussed herein, enables the computer system 80 to implement a power plant control product and / or method. The computer system 80 can use any solution to obtain power plant control data. For example, the computer system 80 can generate power plant control data and / or is used to generate power plant control data and generate power from one or more data stores, repositories, or sources. It is possible to read out plant control data and receive power plant control data from another system or device inside or outside, such as a power plant, a plant controller, a component controller. In another embodiment, the invention provides a method of providing a copy of a program code, for example for a power plant control program, wherein the program code comprises a process, some or all of which are described herein. It can be implemented. It is understood that aspects of the present invention may be implemented as part of a business method of implementing the processes described herein on a subscription, advertising, and / or fee basis. The service provider may offer to implement a power plant control program and / or method as described herein. In this case, the service provider manages (eg, creates, maintains, etc.) a computer system, such as computer system 80, that implements the processes described herein for one or more customers. , Support, etc.) is possible.

  A computer model of the power plant can be built and then used to control and optimize power plant operation. Such plant models can be dynamic, through ongoing comparison between actual (ie, measured) operating parameters and the same parameters predicted by the plant model. , Can be updated repeatedly. In preparing and maintaining such models, instructions that instruct the processor 82 of the computer system 80 to generate a library of energy system power generation units and components ("library of components") in response to user input. Written or otherwise provided. In some configurations, the user input and generated libraries include characteristics of components that comprise the libraries, and rules that cause scripts to be generated according to behavioral and characteristic values. These property values can be compiled from data stored locally in memory 84 and / or obtained from a central data repository maintained at a remote location. The library of components can include non-physical components, such as economic or legal components. Examples of economic components are the purchase and sale of fuels, examples of legal components are emission limits and credits. These non-physical components can be modeled using mathematical rules, just as the components that represent physical devices can be modeled using mathematical rules. The instructions can be configured to assemble a configuration of energy system components from the library, as can be configured by an operator. A library of energy system components may be provided to allow the user to select components from which to duplicate an actual power plant or create a virtual power plant. Each component can have some characteristics that can be used by the user to enter specific values that match the operating conditions of the actual or virtual power plant being modeled. It will be recognized. Scripts may be generated for the assembled energy system components and their configurations. The generated scripts, when used in an energy system component configuration, include mathematical and / or legal components, mathematical relationships within and / or between energy system components. Can be included. Computer system 80 can then solve the mathematical relationship and show the solved result on display 81. In configurations where the signal may be transmitted from computer 80, the signal may be used to control the energy system according to the solved result. Otherwise, the results may be displayed or printed, and the results may be used to set physical equipment parameters and / or non-physical parameters such as fuel purchase and / or sale. , And / or using the determined non-physical parameters, and thus a suitable or optimized mode of operation is achieved. A library of plant components can include a central repository of data, where a central repository of data stores data related to how each plant component behaves under different parameters and conditions. Represents ongoing accumulation. A central repository of data can be used, for example, to provide "plug data" when sensor data is determined to be unreliable.

  A look at Figures 7 to 9 provides a more detailed discussion of the economic power supply process, which is the case for the control system discussed above, whichever the case may be. Including the schemes that can be used to optimize such power delivery procedures, both from the perspective of the power system central authority participating in or of the individual power plant. From the point of view of the central authority power dispatch, the goal of an economical power supply process is to dynamically respond to changing variables, including changing load requirements or ambient conditions, while still minimizing costs incurred in the system. It will be recognized that this is the case. It will be appreciated that with respect to participating power plants, it is generally recognized that the goal is to utilize available capacity while minimizing costs incurred and maximizing economic benefit. Given the complexity of the power system, the process of economical power supply typically involves frequently adjusting the load of the participating power plants by the power dispatching office. When successful, the process results in available power plants being operated at loads where their incremental power costs are approximately the same, which results in minimizing power costs. However, it also complies with system constraints such as maximum and minimum allowable load, system stability, etc. It will be appreciated that accurate incremental cost data is needed for economical power supply to function optimally. Such incremental cost data has major components including fuel cost and incremental fuel consumption. Incremental fuel consumption data is usually given as a curve of incremental heat rate vs. power output. Specifically, the incremental heat rate IHR of a thermal power unit is defined as the slope of the heat rate curve, where the heat rate of the unit is the ratio of heat input plotted against electrical output at any load. This data error will result in powering the unit at loads that do not minimize the total power generation cost.

  Multiple items can introduce errors into the incremental heat rate curve. These can be grouped into two categories. The first category includes items that create errors that exist when the data is provided to the power dispatching office. For example, if the data is collected by testing, the error due to the inaccuracy of the instrument will be included in all calculations made with it. As discussed in more detail below, certain aspects of the present invention provide a method for verifying sensor accuracy during data collection, and the data collected may be unreliable due to sensor malfunction. There is a method for identifying an instance in a timely manner. The second category of errors includes items that cause the data to become inaccurate over time. For example, if the performance of a power generation unit changes due to equipment degradation, or repairs, or changes in ambient conditions, the incremental heat rate data used for powering will be updated with such data. Up to this, an error will occur. One aspect of the present invention is to identify those parameters of a thermal power unit that can significantly affect the incremental heat rate calculation. Knowledge of such parameters, and their relative importance, is then used to determine how often the feed data should be updated to reflect true plant performance. Can be done.

Incremental heat rate data errors lead to situations where the power plant is being falsely powered, which typically results in increased power generation costs for the power system. For example, referring to the graph of FIG. 7, a situation is provided in which the true incremental heat rate differs from the incremental heat rate used in the power feed process. When powering the unit, the power dispatching authority uses the incremental heat rate data in error by "E" as shown. (FIG. 7 assumes that the incremental heat rate of the power system is unaffected by the load imposed on a given unit, which the power system compares to the size of a given power generating unit. If it is larger, the it should be noted that obtained substantially correct.) are shown as the power generation unit becomes a to be fed at L _1, L ₁ is utilized Based on the possible information, the unit and system incremental heat rates are equal loads. If the correct incremental heat rate information is used, the unit is must have been fed at L _2, L ₂ is the load true incremental heat rate of the plant is equal to the incremental heat rate of the power system. As will be appreciated, the error results in underutilization of the power plant. In another case where it applies, i.e., where the position of the correct incremental heat rate plot is opposite the true incremental heat rate plot, the error results in the unit being over-allocated. , It may require the unit to operate inefficiently in order to meet its powered load start-up schedule. From the point of view of the central feeding command authority of the power system, reducing errors in the data used during the feeding process reduces the total system fuel cost, increases system efficiency and / or load. It will be appreciated that this will reduce the risk of not meeting the requirements. For power plant operators in the system, reducing such errors should promote full utilization of the plant and improve economic benefits.

  8 and 9 illustrate a schematic diagram of a plant controller 22 and a flow diagram 169 of a control method, respectively, in accordance with aspects of the present invention. In these examples, a method is provided that illustrates economical optimization within a power system that uses economical power distribution and distributes load among potential providers. The basic process of economical power supply is a process that can be used in various ways between any two levels defined in a multi-layered hierarchy common to many power systems. In some cases, for example, an economic power supply process may be used as part of a competitive process, which causes a central government authority or industry co-operative to divide the load among several competitors. Alternatively, the same principles of economical power supply can be used to share the load between co-owned power plants, to minimize power generation costs for the plant owners. This principle can also be used at the plant level so that the operator or plant controller can share its load requirements among the different local power generation units available. It will be appreciated that unless otherwise stated, the systems and methods of the present invention are generally applicable to any of these possible manifestations of an economical power supply process.

  In general, the power supply process seeks to minimize power generation costs in the power system through the generation of power supply schedules, in which the incremental power generation costs for each participating power plant or unit are approximately the same. As will be appreciated, a number of terms are often used to describe the economical feeding process and will therefore be defined as follows. The “prediction plan target period” is a predetermined period during which optimization will be performed. For example, a typical forecast planning horizon can be from hours to days. The "interval" in the forecast plan period is the predetermined time resolution of optimization, that is, the above-mentioned "optimization cycle", which is how often the optimization is performed during the forecast plan period. Describe what will be done. For example, typical time intervals for optimization cycles can be from seconds to minutes. Finally, the "predicted length" is the number of time intervals during which the optimization will be carried out, which can be obtained by dividing the forecast planning period by the time intervals. Thus, for a 12-hour forecasted horizon and a 5 minute time interval, the predicted length is 144 time intervals.

  Aspects of the present invention provide methods and / or controllers of control for power plants and methods and systems for optimizing performance, cost effectiveness, and efficiency. For example, in accordance with the present invention, a minimum variable operating cost may be achieved for a thermal power generation unit or power plant that includes variable performance characteristics and cost parameters (ie, fuel cost, ambient conditions, market conditions, etc.) and life cycle costs. (Ie variable operations and their impact on maintenance schedules, parts replacement, etc.). By varying one or more parameters of the thermal power generation unit taking such factors into account, it is possible to obtain more economic benefits from the unit over its useful life. For example, in power plants including gas turbines, combustion temperatures provide a more economically desirable load level based on operating profiles, ambient conditions, market conditions, expectations, power plant performance, and / or other factors. Can be changed. As a result, waste of parts with a time-based remaining life remaining in a unit with a limited number of starts may be reduced. In addition, a power plant control system that includes a feedback loop updated with substantially real-time data from sensors that are regularly tested and confirmed to be working correctly, allows for further plant optimization. Become. That is, according to a particular embodiment of the present invention, by introducing a real-time feedback loop between the power plant control system and the power supply commanding authority, the target load and unit start-up plan is made into a very accurate offer curve. The offer curve is constructed based on real-time engine performance parameters.

  FIG. 8 illustrates a schematic design of an exemplary plant controller 22 according to aspects of the invention. It will be appreciated that the plant controller 22 may be particularly suitable for implementing the method 169 of FIG. Due to this, FIGS. 8 and 9 will be discussed together, but it is recognized that each can have aspects applicable to more general usage situations. It will be. The power system 10 shown in FIG. 8 includes a "power plant 12a" as well as "another power plant 12b", a plant controller 22 is dedicated to the "power plant 12a", and "another power plant 12b" is , A power plant in a power system that competes with the power plant 12a. Also, as shown, the power system 10 includes a power supply command authority 24, which, via a dedicated system controller 25, supplies power between all participating power plants 12a, 12b in the system. Manage the process.

  The power plant 12a may include a number of sensors 46 and actuators 47 by which the plant controller 22 monitors operating conditions and controls the operation of the plant. The plant controller 22 can communicate with multiple data resources 26, which can be located remotely from the plant controller 22, accessible by the communication network, and / or locally. It may be contained and accessible by the local network. As shown, the schematic of the plant controller 22 includes several subsystems separated from each other by several boxes. These subsystems or "boxes" are primarily functionally separated and intended to aid in explanation. However, unless otherwise stated, isolated boxes may or may not represent individual chips or processors or other individual hardware elements, and may be implemented in the plant controller. It will be appreciated that separate sections of the computer program code that are represented may or may not represent separate sections. Similarly, the method 169 is divided into two main sections or blocks, for convenience and for explanation purposes only. Any or all of the separate boxes shown in FIG. 8, as well as any or all of the separate blocks or steps shown in FIG. It will be appreciated that they can be combined into

  The method 169 of FIG. 9 can begin, for example, with the control section 170, which receives or gathers current information and data for use (at step 171), the data being market data. , Motion data, and / or ambient data. Within the plant controller 22, the corresponding control module 110 may be arranged to request / receive this type of data from the data resource 26 or any other suitable source. The control module 110 may also be configured to receive a target load 128 from the power dispatching authority 24 (but during the initial run, such target load may not be available and a predetermined initial target load may be used. ). Ambient data may be received from remote or local data repositories and / or predictive services and may be included as a component of data resource 26. Ambient data may also be gathered via ambient sensors located around the power plant 12a and may be received via a communication link with the power dispatching authority 24. In accordance with aspects of the present invention, ambient data includes historical data describing ambient conditions for power plant 12a, current data, and / or expected data, including, for example, air temperature, relative humidity, pressure, and the like. It is possible. Market data may be received from remote or local data repositories and / or forecasting services and may be included as a component of data resource 26. Market data may also be received via a communication link with the power dispatching authority 24. In accordance with aspects of the present invention, market data includes historical data, current data, and / or forecast data that describes market conditions for power plant 12a, such as energy selling prices, fuel costs, labor costs, etc. including. Also, operational data may be received from a data repository and / or anticipation service and may be included as a component of data resource 26. The operational data can include data collected from a plurality of sensors 46, which are deployed within the power plant 12 and its plant components 49 and are physically associated with plant operation. Measure various parameters. The operational data can include historical data, current data, and / or expected data, as well as various process inputs and outputs.

  As can be seen in FIG. 9, the initial set point for the power plant 12 can be determined, for example, by the control model 111 in the plant controller 22 of FIG. For example, the control model 111 uses the thermodynamic and / or physical details of the power plant 12, as well as additional information such as ambient data, or market data, or process data (see FIG. 9). It may be configured (at step 172) to determine values of operating parameters for the power plant 12. In some cases, for example, the value of the operating parameter can be the value that will be needed to achieve sufficient power output to meet the target load. The determined value may be used (also in step 172 of FIG. 9) as an initial set point for each operating parameter of the power plant 12. It is recognized that examples of such operating parameters can include fuel flow rate, combustion temperature, position with respect to inlet guide vanes (if guide vanes are present), steam pressure, steam temperature, and steam flow rate. The Rukoto. The performance indicators may then be determined (in step 173 of FIG. 9) using the performance model 112 of the plant controller 22. Performance indicators can provide operational characteristics such as efficiency of power plant 12. Performance model 112 may be configured to use thermodynamic and / or physical details of power plant 12 and set points determined by control model 111 to determine values of operating characteristics of power plant 12. It is supposed to do. Performance model 112 may be configured to take into account additional information such as ambient conditions, market conditions, process conditions, and / or other relevant information.

  In addition, according to certain aspects of the present invention, the power plant 12 (at step 174 of FIG. 9) may be used, for example, using the life cycle cost (LCC) model 113 included in the plant controller 22 of FIG. An estimate of the LCC of the can be determined. The LCC model 113 can be a computer program or the like, which uses physical and / or cost information about the power plant 12 and set points from the control model 111 to generate electricity. The plant 12 may be configured to determine an estimated life cycle cost. Life cycle costs can include, for example, the total cost, maintenance costs, and / or operating costs of the power plant 12 over its useful life. LCC model 113 may be configured to consider the results of performance model 112 for enhanced accuracy. Therefore, the LCC model 113 uses the determined set points of the control model 111 and the operating characteristics from the performance model 112, as well as other information, to estimate the useful life of the power plant 12 as desired. , It can be estimated how much it can cost to operate and / or maintain the power plant 12 during its useful life. As mentioned above, the useful life of a power plant may be expressed in terms of hours of operation and / or number of starts, and a given power plant has an expected useful life that can be provided by the manufacturer of the power plant. Therefore, the predetermined value of the expected useful life may be used at least as a starting point for the LCC model 113 and / or the enhancement module 114.

  Using information from other embodiments of the invention, such as initialization points, performance indicators, and results from determining an estimated life cycle, etc., as described below (step 175 At) the optimization problem for the power plant 12 may be solved. Such an optimization problem may include multiple equations and variables depending on the depth of analysis desired, and may include an objective function, which in embodiments is , LCC-based objective function. The solution can include providing enhancements or enhancements to the operating parameters of the power plant 12, such as by minimizing an LCC-based objective function (also step 175). In an embodiment, the optimization problem solution may be implemented by the enhancement module 114 of the plant controller 22 of FIG.

  As is known from optimization theory, the objective function represents the property or parameter to be optimized, and depending on how the optimization problem is defined, many variables and / or parameters Can be taken into account. In optimization problems, the objective function may be maximized or minimized, depending on the particular problem and / or the parameters represented by the objective function. For example, as shown above, the objective function representing the LCC according to the embodiment may include at least one operating parameter that may be used to run the power plant 12 to keep the LCC as low as possible. Will be minimized to produce The optimization problem for the power plant 12, or at least the objective function, is power plant characteristics, site parameters, customer specifications, results from the control model 111, performance model 112 and / or LCC model 113, ambient conditions, market conditions. , And / or process conditions, and any additional information, etc. that may be suitable and / or desired, and the like can be taken into account. Such factors may be collected into the objective function terms, eg, an LCC-based objective function is intended to include maintenance and operating costs expressed over time, where time is the estimated component life. It is the period covered by the forecast plan based on the number of years. Complex objective functions and / or optimization problems may be used in implementations of the invention, each of which may include many or all of the various functions and / or factors described herein. It will be recognized that there is.

  For example, maintenance costs may be determined by modeling parts of the power plant 12 and estimating wear based on various parameters such as those already discussed. It will be appreciated that any component of the power plant 12 may be modeled for these purposes. However, in a practical application, parts associated with a smaller number of larger parts of the power plant 12, or a smaller number of selected parts may be modeled and / or instead of modeling, for some parts, A constant or plug value can be used. Whatever level of detail is used, such LCC-based objective function minimization is part of the optimization problem, which is much like that provided above. As a result of factors of, may include at least one enhanced or augmented operating parameter of the power plant 12 that may vary for a given power plant and, for example, according to minimizing LCC, etc. It is possible. In addition, predetermined uptime and / or downtime, predetermined upper and / or lower temperatures at various locations in the power plant 12, predetermined torque, predetermined power output, and / or as desired and / or Or one of ordinary skill in the art will recognize that at least one constraint may be imposed on the optimization problem, such as other constraints as appropriate. Unless otherwise stated, it is important to determine what constraints should be applied and in what manner for a given optimization problem. Within the management authority of the vendor. Moreover, one of ordinary skill in the art will recognize situations in which additional optimization theory techniques may be applied, such as adding slack variables to enable feasible solutions to optimization problems.

  Known techniques may be used to solve optimization problems with the operation of the power plant 12, for example, by the enhancement module 114 (FIG. 8). For example, integer programming, linear methods, mixed integer linear methods, mixed integer nonlinear methods, and / or other techniques may be used as needed and / or desired. In addition, as can be seen in the exemplary objective function, the optimization problem can be solved over the forecasted time horizon to provide an array of values for at least one operating parameter of the power plant 12. is there. The enrichment or augmentation may be performed over a relatively short forecasted time horizon, such as, for example, on the order of 24 hours or even minutes, but the enrichment module 114 (FIG. 8) will depend on the depth of analysis desired. Thus, it is possible to use a longer forecast planning period, for example up to the estimated useful life of the power plant 12. In an embodiment, the initialization points determined, for example, by the control model 111 (FIG. 8) are adjusted and reinforced in response to and / or as part of the optimization problem solution. It is possible to produce an enhanced, or enhanced, or an optimized set point. In addition, determining an initial set point (in steps 172-175 of FIG. 9) to improve results and / or enhance or enhance the control set point of the power plant 12, the value of the performance index. Iterative can be used in conjunction with determining, estimating the estimated LCC cost, and enhancing or enhancing.

  As will be described, the offer curve section 180 can generate an offer curve, or set of offer curves, an example of which was shown above in connection with FIG. At plant controller 22, control information 115 from control module 110 and / or data resource 26 may be received by offer curve module 120 (at step 181 of FIG. 9). According to particular embodiments, control information 115 includes control set points, performance, ambient conditions, and / or market conditions. This information may also be known as "as run" information. In addition, ambient condition forecasts 121 and / or market condition forecasts 122 may be received (at step 182). According to particular embodiments, a database 123 may be included, which may store current information, “on the run” information, and / or historical information locally, which may be Includes any or all of conditions, market conditions, power plant performance information, offer curves, control set points, and / or any other information that may be appropriate. The database 123 may be used to provide information for simulating the operation of the power plant 12 (at step 183) using, for example, an off-line model 124 of the power plant 12.

  The off-line model 124 can include a model similar to the control model 111, but can also include additional modeling information. For example, the off-line model 124 may incorporate some or all of the control model 111, the performance model 112, the LCC model 113, and / or additional modeling information. By running the offline model 124 with the setpoints and / or information from enhancing or augmenting the LCC, the output of the offline model 124 outputs power for each time interval during the forecasted time horizon. It may be used to determine an estimate for the cost of production and for various values of the power output of the power plant 12, generating (in step 184) one or more offer curves 125. Then, one or more offer curves 125 may be sent (at step 185) to the power dispatching authority 24 or otherwise provided. The offline model 124 may use any suitable information, such as historical information, current information, and / or forecast information, in determining the estimated operating costs and / or conditions for the power plant 12. It is possible. In addition, the offline model 124 in embodiments may be tuned (at step 186), such as by the model tuning module 126. Tuning may, for example, periodically adjust parameters for the off-line model 124 based on information received and / or provided by other components of the plant controller 22 to improve the actual operation of the power plant 12. Can be included to better simulate the operation of the power plant 12. Thus, for a given set of operating parameters, if the plant controller 22 observes actual process conditions that are different from those predicted by the off-line model 124, the plant controller 22 will accordingly generate the off-line model 124. It can be changed.

  In addition to the offer curve 125 from the power plant 12a, as illustrated, the power dispatching authority 24 may receive the offer curve 125 from another power plant 12b under its control. The power dispatching authority 24 can access the offer curve 125 and generate power dispatch schedules to accommodate the load on the power system 10. The power dispatching authority 24 may additionally take into account expected ambient conditions, load expectations, and / or other information as needed and / or desired, which may include power dispatching It may be received from various local or remote data resources 26 accessed by the authorities 24. As shown, the power delivery schedule produced by the power dispatching authority 24 includes control signals for the power plant 12 including the target load 128, and the plant controller 22 controls its control as described above. It is possible to respond to the signal.

  Inclusion of life cycle cost considerations, as described herein, increases the scope and accuracy of the plant model used in the optimization process and, in doing so, enhances the procedure. It will be appreciated that it is possible to play a role in enabling. Offer curve 125 may represent a variable cost (measured in dollars per megawatt hour for megawatt power plant output), as described above. Offer curve 125 may include an incremental variable cost offer curve and an average variable cost offer curve. As can be seen, embodiments of the present invention are able to provide a variable cost accurate assessment via their generated offer curves 125. Using the embodiments of the present invention, the incremental variable cost offer curve is shown to very closely predict the actual incremental variable cost curve, while the average variable cost offer curve is the actual average. It is shown to predict variable cost curves very closely. The accuracy of the offer curve generated by embodiments of the present invention shows that the various models used in the plant controller 22 of FIG. 8 provide representative models suitable for review purposes. ing.

  10-12, another aspect of the present invention is described with reference to and including the particular systems and methods provided above. FIG. 10 is a dataflow diagram demonstrating an architecture for a plant optimization system 200 that may be used in a combined cycle power plant having gas and steam turbine systems. In the provided embodiment, the system 200 includes a monitoring and control instrument 202 associated with each of the gas turbine (202) and steam turbine system (204), such as the sensors and actuators discussed above, and the like. Including 204. Each of the monitoring and control instruments 202, 204 can send a signal to the plant controller 208 that is indicative of the measured operating parameter. The plant controller 208 receives the signals, processes the signals according to a predetermined algorithm, and sends control signals to the monitoring and control instruments 202, 204 to affect changes to plant operation.

  The plant controller 208 interfaces with the data collection module 210. The data collection model 210 may be communicatively coupled to a database / historian 212 that maintains archived data for future reference and analysis. The heat balance module 214 is capable of receiving data from the data collection model 210 and the database / historian 212 on demand, and in order to fit the measured data as closely as possible to the mass of the power plant. Handles algorithms for tuning balance and energy balance models. Discrepancies between the model and the measured data can indicate errors in the data. As will be appreciated, the performance module 216 can use the plant equipment model to predict expected performance of key plant components and equipment. The difference between expected performance and current performance can be indicative of a deterioration in conditions of plant equipment, parts, and components such as, but not limited to, fouling, scaling corrosion, and breakage. In accordance with aspects of the present invention, the performance module 216 can track degradation over time so that the performance issues that have the most significant impact on plant performance are identified.

  An optimizer module 218 may be included as shown. The optimizer module 218 can include methodologies for optimizing the economical power supply of the plant. For example, according to an embodiment, a power plant may be powered based on a heat rate or an incremental heat rate according to the assumption that the heat rate is equivalent to a financial source. The power plant includes an additional manufacturing process (not shown) for which steam is used directly (ie, the steam produced is separated from power generation in the steam turbine). In an alternative scenario, the optimizer module 218 may be able to solve the optimization problem and components with higher heat rates may be powered. It will be. For example, in certain situations, the demand for steam may exceed the demand for electricity, or the electrical output may be constrained by electrical system requirements. In such cases, powering a lower efficiency gas turbine engine may allow more heat to be recovered without raising the electrical output beyond limits. In such a scenario, powering components with higher heat rates is an economically optimized alternative.

  The optimizer module 218 may be selectable between online (automatic) mode and offline (manual) mode. In online mode, the optimizer 218 determines the cost of electricity generated, the incremental cost at each level of power generation, the cost of process steam, and a predetermined periodicity, eg, in real time or once every 5 minutes. Automatically calculate current plant economic parameters such as plant operating profits at Offline mode simulates steady-state performance, analyzes “what-if” scenarios, budgets and upgrade options, and present power generation capacity, target heat rates, and current plant operation to guarantee conditions. It can be used to predict the effects of corrections, operational constraints and maintenance effects, and fuel consumption. The optimizer 218 does not calculate output based on the efficiency of combining plant heat balance and plant financial models, but rather on the basis of real-time economic cost data, output prices, load levels, and equipment degradation. Compute the profit-optimized output for the plant. The optimizer 218 may be tuned to individually adapt to the degradation of each component, may produce an advisory output 220, and / or may produce a closed feedback loop control output 222. The advisory output 220 is designed to encourage the operator where to set the controllable parameters of the power plant, to optimize each plant component and maximize profitability. In the exemplary embodiment, advisory output 220 is a computer display screen communicatively coupled to computer-implemented optimizer module 218. In an alternative embodiment, the advisory output is a remote workstation display screen and the workstation accesses the optimizer module 218 over the network. The closed feedback loop control output 222 is capable of receiving data from the optimizer module 218 and provides optimized set points and / or bias set points for the modules of system 200 to implement real time feedback control. calculate.

  FIG. 11 is a simplified block diagram of a real-time thermal power plant optimization system 230, which according to aspects of the present invention includes a server system 231 and a plurality of client subsystems. And a plurality of client subsystems, also referred to as client system 234, are communicatively coupled to the server system 231. As used herein, real time refers to a result, eg, a computer calculation, that occurs in a substantially short period of time after a change in input that affects the result. The time period represents the amount of time between each iteration of the task that is periodically repeated. Such recurring tasks may be referred to herein as periodic tasks or cycles. The time period is a design parameter of the real-time system, which may be selected based on the importance of the result and / or the system's ability to implement processing of the inputs to generate the result. Additionally, events that occur in real time occur without substantial intentional delay. In an exemplary embodiment, the calculations may be updated in real time with a periodicity of 1 minute or less. Client system 234 can be a computer including a web browser such that server system 231 can access client system 234 via the Internet or some other network. Client system 234 can be interconnected to the Internet through many interfaces. Client system 234 can be any device capable of interconnecting to the Internet. As described in more detail below, the database server 236 is connected to a database 239 that contains information about a number of things. In one embodiment, centralized database 239 includes aspects of data resource 26 discussed above, centralized database 239 is stored on server system 231, and also through client system 234 to server system 231. By logging on, it may be accessed by a potential user on one of the client systems 234. In an alternative embodiment, the database 239 may be stored remotely from the server system 231 and decentralized.

  In accordance with aspects of the present invention, some of the control methods discussed above can be developed for use with the system diagrams of FIGS. 10 and 11. For example, one method involves simulating power plant performance using a plant performance module in a software code segment that receives power plant monitoring instrument data. Data may be received over the network from a plant controller or database / historian software program running on a server. Any additional plant components, such as an intake conditioning system or a HRSG duct firing system, etc., can be simulated in a manner similar to those used to simulate power plant performance. Determining the performance of each plant component in the same manner allows the overall power plant to be treated as a single plant, determining such setpoints for each component separately. Rather, determine an optimized set point for the power plant. The measurable quantity for each plant component may be parameterized to represent the output or power plant efficiency for each component. Parameterizing plant equipment and plant performance includes, but is not limited to, gas turbine compressors, gas turbines, heat recovery steam generators (HRSG), draft fans, cooling towers, condensers, feed water heaters, evaporation. Includes calculating efficiency for components such as vessels, flash tanks, etc. Similarly, heat rate and performance calculations can be parameterized, and the resulting system of equations can be solved in real-time, and the calculated results are available without intentional delay from the time each parameter is sampled. It will be recognized that it is becoming possible. Solving parameterized simultaneous equations and constraints also determines current heat balance for power plants, including, but not limited to, operating reserve requirements, electrical system demand, and maintenance activities. , Determining the expected performance using current constraints on the operation of the power plant, such as, clean water demand, and component outages. Solving the parameterized equations and constraints also determines the parameters to adjust to modify the current heat balance so that future heat balance equals the determined expected performance. Can be included. In an alternative embodiment, solving the parameterized simultaneous equations and constraints determines the inlet condition to the power plant, the power plant based on the determined inlet condition and a predetermined model of the power plant. Predicting output, determining the current output of the power plant, comparing the predicted output with the determined output, and adjusting plant parameters until the determined output equals the predicted output Including doing. In an exemplary embodiment, the method also uses parameterized equations to correlate controllable plant parameters, plant equipment, and plant performance, and to minimize the heat rate of a power plant. And / or defining objectives for optimization using objective functions, including maximizing the benefits of the power plant, and using constraints to physically control the behavior of each individual piece of equipment. Possible ranges and / or defining overall limits, which includes maximum power production, maximum fuel consumption, and the like.

  FIG. 12 is a flowchart of an exemplary method 250 for solving parameterized simultaneous equations and constraints in accordance with the present invention. The method 250 determines (at 252) a current heat balance for the power plant, (at 254) uses the current constraints for operation to determine expected performance, and (at 256) a future heat balance. Determining the parameters to adjust to modify the current heat balance so that is equal to the determined expected performance. The method 250 also determines 258 inlet conditions to the power plant, predicts 260 the power plant output based on the determined inlet conditions and a predetermined model of the power plant, and determines the current power plant current. Determining the output 262, comparing the predicted output with the determined output 264, and adjusting plant parameters 266 until the determined output is equal to the predicted output. It is noted that the method described and the system discussed in connection with FIGS. 10 and 11 provide a cost effective and reliable means for optimizing a combined cycle power plant. Will be recognized.

  Turning now to Figures 13-16, attention will be paid to several flow diagrams and system configurations illustrating a control methodology in accordance with certain aspects of the present invention. Generally, according to an exemplary embodiment, a thermal power generation unit, such as a gas turbine system, or a control system for a power plant, may include first and second instances of the model, and The first and second instances of model the behavior of the turbine, for example, by utilizing a physics-based model or by mathematically modeling (eg, a transfer function, etc.). The first model, which may also be referred to as the "first-order model", may provide the current operating parameters of the gas turbine system, which determines the mode of operation of the turbine and its corresponding operating conditions. Describe. As used herein, “parameters” include, but are not limited to, temperature, pressure, gas flow, and compressor, combustor, and turbine efficiency at defined locations within the turbine. Represents items that can be used to define turbine operating conditions, such as levels. Performance parameters may also be referred to as "model correction factors" and represent the factors used to adjust the first or second model to reflect the behavior of the turbine. Inputs to the first model can be sensed or measured and provided by an operator. In addition to current performance parameters, the method of the present invention receives or otherwise receives information about external factors or disturbance variables, such as ambient conditions, that may affect current or future operation of the gas turbine system. Can be obtained.

  The second model (also referred to as the "quadratic model" or "predictive model") takes into account current operating parameters, such as manipulated variables, and one or more disturbance variables, Generated to identify or predict one or more operating parameters of the gas turbine system, such as controlled variables. Exemplary operating parameters of the turbine include, but are not limited to, actual turbine operating conditions such as exhaust temperature, turbine power output, compressor pressure ratio, heat rate, emissions, fuel consumption, expected revenue, etc. . Thus, this second or predictive model may be utilized to indicate or predict turbine behavior at a particular operating set point, performance goals, or operating conditions that differ from current operating conditions. As used herein, the term "model" generally refers to the act of modeling, simulating, predicting, or exhibiting based on the output of a model. Although the term "second model" is utilized herein, in some cases there may be no difference between the formulation of the first and second models, and "second model" It will be appreciated that the "model of" is adapted to represent the running of the first model by means of adjusted parameters or additional or different inputs.

  Thus, by modeling a turbine operating behavior utilizing a second model or predictive model that takes into account external factors and / or different operating conditions, turbine control may be performed under these different operating conditions or It can be adjusted to operate more efficiently, taking into account external factors outside. Therefore, the system allows for automated turbine control based on modeled behavior and operating characteristics. In addition, the modeling system described is capable of generating operator-specific scenarios, inputs, operating points, operating goals, and / or operating conditions and predicting turbine behavior and operating characteristics under these operator-specific conditions. To enable. Predicting such virtual scenarios allows the operator to make more informed controls and operational decisions, such as scheduling, loading, turndown, and so on. As used herein, the term "operating point" generally refers to an operating point, condition, and / or goal and is not intended to be limiting. Thus, operating points can represent goals or set points, such as base load, turndown points, peak fires, and the like.

  One exemplary use of the described turbine modeling system involves adjusting turbine operation to meet grid compliance requirements while still operating at the most efficient level. For example, local grid authorities typically specify requirements that power plants must be able to support the grid during frequency disruptions. Supporting the grid during a disruption involves increasing or decreasing the turbine load under certain conditions, depending on grid conditions. For example, during a turmoil, a power plant is expected to increase its power output (eg, by as much as 2%) to compensate for other supply shortages. Therefore, turbine operation typically constrains the base load point, allowing the turbine to be operated at a power level with margin (also referred to as "preliminary margin") and, if necessary, overload. Increased loads can now be provided without incurring the additional maintenance factors associated with firing. As one example, the spare margin can be 98% of a typical base load, thus allowing the load to be increased to meet grid requirements without exceeding 100% base load. (For example, increase by 2%). However, unexpected external factors such as temperature, humidity, or pressure can adversely affect turbine efficiency. As the daytime temperature rises, the heat causes the turbine to operate at a lower efficiency and the turbine cannot reach the 100% load it was originally planned for, so the turbine needs to operate at the 2% required by the turbine. May not have reserve capacity. To compensate for that, the traditional heat rate curve causes the turbine to operate in a more efficient state throughout the day, taking into account possible mechanical efficiency losses (eg, at 96%, etc.). . However, the turbine modeling system described herein allows modeling turbine behavior in real time according to current external factors (eg, temperature, humidity, pressure, etc.), and thus the current ambient. Allows control of turbine operation to operate most efficiently given the conditions. Similarly, future turbine behavior may be predicted, such as predicting turbine behavior in response to daily thermal fluctuations, and turbine operation planning may provide the most efficient and economically profitable operation. enable. As another example, a power plant typically makes a decision as to whether to shut down a gas turbine at night or simply reduce the power level (eg, turn down). Turbine operating characteristics such as emissions, exhaust temperature, etc. influence this decision. The turbine modeling system described herein can be utilized to make more intelligent decisions, either in advance or in real time or near real time. External factors and expected turbine operating parameters can be fed into the second model to determine what the turbine operating characteristics will be. Therefore, taking these characteristics (eg efficiency, emissions, costs, etc.) into account, the modeled characteristics can be utilized to determine whether the turbine should be shut down or turned down. Is.

  As yet another example, the turbine modeling system may be utilized to assess the benefits of performing turbine maintenance at a given time. The turbine modeling system of the present invention may be utilized to model turbine operating characteristics at current capabilities based on current performance parameters. Operator-specific scenarios can then be generated that model the operating characteristics of the turbine when maintenance is performed (eg, improve performance parameter values to indicate expected increased performance). For example, as the turbine deteriorates over time, performance parameters reflect mechanical deterioration. In some cases, maintenance may be performed to improve their performance parameters and thus the operating characteristics of the turbine. By modeling or predicting the improved operating characteristics, a cost-benefit analysis can be performed to compare the cost with the benefit obtained by performing maintenance.

  FIG. 13 illustrates an exemplary system 300 that may be used to model turbine operating behavior. According to this embodiment, a power plant 302 is provided that includes a gas turbine having a compressor and a combustor. The intake duct to the compressor feeds ambient air and possibly injected water to the compressor. The configuration of the intake duct contributes to the pressure loss of the ambient air flowing into the compressor. The exhaust duct for the power plant 302 directs combustion gases from the outlet of the power plant 302, eg, through emission control and sound absorbing devices. The amount of inlet pressure loss and back pressure can change over time due to the addition of components to the intake and exhaust ducts and due to clogging of the intake and exhaust ducts.

  The operation of power plant 302 may be monitored by one or more sensors that detect one or more observable conditions of power plant 302, ie, operating or performance parameters. In addition, external factors such as the ambient environment can be measured by one or more sensors. In many cases, duplexed or triplicated sensors are capable of measuring the same parameters. For example, a group of redundant temperature sensors can monitor ambient temperature surrounding power plant 302, compressor discharge temperature, turbine exhaust gas temperature, as well as other temperatures throughout power plant 302. Similarly, a group of redundant pressure sensors can monitor ambient pressure and static and dynamic pressure levels at the compressor inlet and outlet, turbine exhaust, and elsewhere throughout the engine. A group of redundant humidity sensors is capable of measuring ambient humidity in the compressor intake duct. The group of redundant sensors may also include flow sensors, speed sensors, flame detector sensors, valve position sensors, guide vane angle sensors, etc. that sense various parameters related to the operation of the power plant 302. The fuel control system can regulate the fuel flowing from the fuel supply to the combustor. The fuel controller can also select the type of fuel for the combustor.

  As stated, "operating parameters" include temperature, pressure, compressor pressure ratio, gas flow at a defined location in the turbine, load set point, combustion temperature, and level of turbine or compressor degradation. And / or represent items that may be used to define operating conditions of a turbine system, such as one or more conditions corresponding to a level of turbine or compressor efficiency. Some parameters are measured directly. Other parameters are estimated by the turbine model or are known indirectly. Still other parameters can represent hypothetical or future conditions and can be defined by the plant operator. The measured and estimated parameters can be used to represent a given turbine operating condition. As used herein, a "performance indicator" is an operating parameter that is derived from the value of a particular measured operating parameter and represents a performance criterion for the operation of a power plant over a defined period of time. There is. For example, the performance index includes heat rate, output level, and the like.

  As shown in FIG. 13, the system 300 includes one or more controllers 303a, 303b, which control the operation of the power plant or power generation unit 302, respectively. Can be a computer system having one or more processors executing programs for. Although FIG. 13 illustrates two controllers, it will be appreciated that a single controller 303 may be provided. According to a preferred embodiment, multiple controllers may be included to provide redundant and / or distributed processing. The control action can depend on, for example, sensor inputs or instructions from a plant operator. The programs executed by controller 303 may include scheduling algorithms, such as algorithms for regulating fuel flow to the combustor, grid compliance, algorithms for managing turndown, and the like. The commands generated by the controller 303 can cause an actuator on the turbine to regulate, for example, a valve between the fuel supply and the combustor, fuel flow, split, and fuel type. Is adjusted. The actuator can adjust the inlet guide vanes above the compressor, or it can activate other control set points above the turbine. It is appreciated that the controller 303, in addition to facilitating control of the power plant, may be used to generate a first model and / or a second model, as described herein. Will be done. The controller 303 can receive operator and / or current modeled output (or any other system output). As described above, the controller 303 can include a memory that stores programmed logic (eg, software) and that also senses operating parameters, modeled. It is possible to store data such as operating parameters, operating limits and goals, operating profiles, etc. A processor can utilize an operating system to execute programmed logic and, in doing so, utilize data stored thereon. A user can interface with the controller 303 via at least one user interface device. The controller 303 is capable of communicating online with the power plant during operation via the I / O interface and offline with the power plant during non-operation. It is possible. One or more of the controllers 303 are capable of implementing the model-based control system described herein, which senses, but is not limited to, operating and performance parameters. Generating, modeling and / or receiving; generating a first power plant model that reflects current turbine operation; sensing, modeling, and / or receiving external factor information; Receiving operator inputs and other variables such as performance objectives; generating a second power plant model that reflects the operation taking into account the additional data provided; current or future turbine operation Controlling; and / or presenting modeled operating characteristics It will be recognized that that can include that. Additionally, it should be appreciated that other external devices, or multiple other power plants or power generation units, can communicate with the controller 303 via the I / O interface. The controller 303 can be remotely located with respect to the power plant it controls. Further, the controller 303 and programmed logic implemented by the controller 303 can include software, hardware, firmware, or any combination thereof.

  The first controller 303a, which can be the same or different controller as the second controller 303b, as described, models the power plant 302 with a first model or a first order model 305. May be operable to model, which includes modeling current performance parameters of the turbine. The second controller 303b may be operable to model turbine operating characteristics under different conditions via the second or predictive model 306. The first model 305 and the second model 306 can each be a configuration of one or more mathematical representations of turbine behavior. Each of these representations is dependent on the input value and is capable of producing a modeled estimate of the operating parameter. In some situations, the mathematical representation can generate surrogate operating parameter values that can be used in situations where measured parameter values are not available. The first model 305 is then based on the current performance parameters of the power plant 302 and any other factors such as external factors, operator supplied commands or conditions, and / or adjusted operating conditions. And can be utilized to provide a basis and / or input to the second model 306 to determine turbine operating characteristics. As explained above, the “second model 306” may simply be an instance of the same model as the first model 305, and the second model 306 may be an external factor, a different behavior. It will be appreciated that additional or different inputs, such as points, etc., may be considered and different inputs may be considered to model different performance parameters or turbine behavior. The system 301 can further include an interface 307.

  Continuing to refer to FIG. 13, a brief description of the interrelationships between system components is provided. As described, the first or first order model 305 models the current performance parameters 308 of the power plant 302. These current performance parameters 308 include, but are not limited to, conditions corresponding to levels of turbine degradation, conditions corresponding to levels of turbine efficiency (eg, heat rate or fuel to power output ratio), inlet guide vanes. Angle, amount of fuel flow, turbine speed, compressor inlet pressure and temperature, compressor outlet pressure and temperature, turbine exhaust temperature, generator power output, compressor air flow, combustor fuel / air ratio, combustion temperature (turbine Inlet), combustor flame temperature, fuel system pressure ratio, as well as acoustic characteristics. Some of these performance parameters 308 may be measured or sensed directly from turbine operation, and some may be modeled based on other measured or sensed parameters. The performance parameter may be provided by the first model 305 and / or may generally be provided by the controller, such as when sensed and / or measured by the controller. Upon generating the first model 305, a performance parameter 308 (which is intended to represent any turbine behavior provided by the model) is provided to generate a second or predictive model 306. To be done. Other variables 309 may be provided to the second model 306 depending on its intended use. For example, other variables can include external factors, such as ambient conditions, which are generally out of control and have to be simply addressed. In addition, other variables 309 are generated by the controller 303, such as a controller-specific scenario or operating point (eg, turbine control based on the first model 305, or otherwise via the controller 303). Provided, for example, a turbine operating point) and the measured input, the measured input possibly being described as being modeled by the first model 305. Could be some or all of the same measured input. As described with reference to FIG. 14, an operator-specific scenario 313 (eg, one or more operator-supplied commands indicating different turbine operating points or conditions) may also be transmitted to the second operator via operator input. Model 306 of For example, as one illustrative use, another variable 309 may include a controller-specific scenario, where the controller-specific scenario is an additional input such as an external factor or a measured input. The second model 306 is provided as one or more inputs when attempting to model current turbine behavior in real-time or near real-time. By utilizing, in addition to one or more of these additional inputs, the controller-specific scenarios of the first model, the expected real-time behavior of the power plant 302 is to account for these additional inputs. Taking into account, it can be modeled by the second model 306, and it can be utilized to control the power plant 302 or to adjust the first model 305 by the control profile input 310.

  Referring to FIG. 14, operator-specific operating modes or scenarios 313 are provided as one or more inputs to a second model or predictive model 306 via interface 307, and then a second model or predictive model. 306 models or predicts future turbine behavior under various conditions. For example, an operator may supply commands to interface 307 to generate scenarios in which power plant 302 operates at different operating points (eg, different loads, configurations, efficiencies, etc.). As an example for illustrative purposes, the set of operating conditions may be provided via an operator-specific scenario 313, which may be delivered the next day (or other future time), such as ambient conditions or demand requirements. Frame) represents the expected conditions for These conditions can then be used by the second model 306 to generate expected or predicted turbine operating characteristics 314 for the power plant 302 during that time frame. When running the second model 306 under an operator-specific scenario, expected operating characteristics 314 include, but are not limited to, base load output capacity, peak output capacity, minimum turndown point, emission level, Represents turbine behavior, such as heat rate. These modeled or predicted operating characteristics 314 may be useful when planning and committing to power production levels, such as for power spot market planning and bidding, for example.

  FIG. 15 illustrates an exemplary method 320 by which embodiments of the invention can operate. A flowchart of the basic operation of a system for modeling a turbine is provided, such as may be performed by one or more controllers, such as those described with reference to FIGS. 13 and 14. The method 320 may begin at step 325, where the controller causes the first or first order model to determine one or more current performance parameters of the turbine according to the current operation. It can be modeled. To generate this first model, the controller can receive, as an input to the model, one or more operating parameters indicative of the current operation of the turbine. As explained above, these operating parameters can be sensed or measured, and / or they can be modeled so that they can occur if the parameters could not be sensed. Current operating parameters can include any parameters indicative of current turbine operation, as described above. It will be appreciated that the methods and systems disclosed herein do not directly depend on whether operating parameters are measured or modeled. The controller can include, for example, a model of the gas turbine that is generated. The model can be a composition of one or more mathematical representations of operating parameters. Each of these representations is dependent on the input value and is capable of producing an estimate of the modeled operating parameter. Mathematical expressions can generate surrogate operating parameter values that can be used in situations where measured parameter values are not available.

  At step 330, the controller may receive or otherwise determine one or more external factors that may affect current and / or future operation. As explained above, these external factors are typically (but need not be) in control, so incorporating their effects into the second model is: Beneficial in producing the desired turbine control profile and / or operating behavior. External factors can include, but are not limited to, ambient temperature, humidity, or atmospheric pressure, as well as fuel composition and / or feed pressure, which can affect turbine operating behavior. These external factors can be measured or sensed, estimated (for example, if the operator requires expected behavior based on virtual scenarios or future conditions), or otherwise May be manually provided by and / or may be provided by a third party information resource (eg, weather service, etc.).

  At step 335, the controller can receive the adjusted operating point and / or other variables to predict turbine behavior at conditions different from the current turbine conditions. Adjusted operating points include, but are not limited to, modeling a turbine at a reserve margin (eg, 98% of base load), or modeling a turbine at, for example, peak load or during turndown. In some cases, it may include specifying a desired output level. Operating points include, but are not limited to, high temperature gas flow path durability (or combustion temperature), exhaust flame durability, NOx emissions, CO emissions, combustor lean blowout, combustion dynamics, compressor surge, It may further include operating limits such as compressor icing, compressor aerodynamic limits, compressor clearance, compressor discharge temperature, and the like. Thus, by providing these adjusted operating points or other variables, the operator is able to provide virtual scenarios, for which the turbine model is Of operating characteristics of the turbine, which may be useful for controlling future operation of the turbine and / or for planning future power generation and shutdown plans.

  Following step 335 is step 340, in which the second model or predictive model of the turbine is the first model generated in step 325 and, optionally, external factors and / or , 335 based on the adjusted operating point or other variable provided in step 335. Thus, this second or predictive model is capable of accurately indicating or predicting operating parameters and, therefrom, performance indicators for the turbine during future operating periods.

  In step 345, the modeled performance may be utilized to adjust current or future turbine operation and / or display the modeled performance to an operator. Thus, when adjusting current turbine operation, the turbine controller may, for example, modify the current control model (e.g. Model) or modeled performance parameters as input to modify the current control profile. This real-time or near real-time control of the turbine will be performed when the input to the second model generated in step 340 is representative of current turbine conditions or current external factors. Is assumed. For example, when the second model represents performance characteristics that take into account temperature, pressure, or humidity, and / or turbine operating parameters or performance parameters that more accurately represent turbine degradation and / or efficiency, the step of Real-time or near real-time adjustments at 345 may be performed. FIG. 16 illustrates one exemplary embodiment that can optionally receive operator-specific inputs to generate expected behavior under different operating conditions. Also, the output of the model generated in step 340 may be displayed or otherwise presented to the operator via the interface. For example, in one embodiment in which the operator provides a virtual operating scenario in step 335, the predicted turbine operating characteristics may be included for analysis and for possible inclusion in future control or planning activities. Can be displayed. Thus, the method 320 may end after step 345, modeling the turbine's current performance parameters with the first model, and then adding additional external factors, adjusted operating points, or otherwise. The same turbine is modeled by taking into account the additional data of the above, and the turbine operation is predicted based on this additional data.

  FIG. 16 illustrates an exemplary method 400 by which alternative embodiments can operate. Provided is an exemplary flowchart of the operation of a system for modeling a turbine, such as may be performed by one or more controllers, such as those described with reference to FIGS. 13 and 14. It Method 400 illustrates the use of system 301, which allows an operator to optionally supply additional variables and utilize modeling capabilities to predict turbine behavior under virtual scenarios. Is. The method 400 may begin at decision step 405, where whether the turbine should be modeled according to current turbine operating and performance parameters, or operator-supplied parameters. Is to be taken into account when generating the model. For example, if the system is being used to predict a virtual operating scenario, the current performance parameters are transferred to the model (assuming the model already reflects basic turbine operation and behavior). May not be needed as input. Therefore, in decision step 405, if it is determined that the current parameter should not be utilized, then operation proceeds to step 410, where the operator provides different performance parameters and different operation. Allows the turbine to be modeled under points and under different operating conditions (eg, under more degraded conditions, at different levels of efficiency, etc.). Otherwise, current performance and / or operating parameters are utilized, eg, as described with reference to step 325 of FIG. 15, and operation continues to step 415. In step 415, the controller may model one or more performance parameters of the turbine according to the operator-supplied input from step 410 or the current operation of the turbine according to a first model or first order model. Is. For example, if the model is generated in step 410 based at least in part on operator-supplied parameters, then the model generated in step 415 produces predicted turbine behavior under those performance parameters. Represent

  Following the step 415 is a decision step 420 where subsequent modeling (eg, a “second model” or a “predictive model”) is performed, such as current temperature, pressure, or humidity. It should be decided whether it should be based on current external factors or on different external factors supplied by the operator. For example, in one scenario, the controller may model turbine operating behavior based on additional data of one or more current external factors, which may be turbine behavior considering current conditions. Will enable further prediction of However, in another scenario, the controller can be utilized to further model the turbine according to conditions supplied by the operator, which allows predicting turbine operating characteristics under various hypothetical scenarios. Thus, in step 420, if it is determined that operator-supplied extrinsic data should be considered when modeling, operation continues to step 425. If not, operation continues to step 430, which utilizes the current external factor. In step 430, the controller determines the external factors that will be considered when generating the second or predictive model, whether they represent the current state or virtual factors. receive. Following step 430 is steps 435-445, which consider different operating points in the same or similar manner as described with reference to steps 325-345 of FIG. , Optionally for generating a predictive model based on the received data and for displaying the predicted behavior, respectively. The method 400 may end after step 445, optionally modeling turbine operating behavior based on operator-specific scenarios.

  Thus, the embodiments described herein utilize turbine models in addition to predicting turbine behavior taking into account current performance parameters and one or more identified external factors, It makes it possible to show the turbine behavior of the actual turbine and the corresponding operating parameters. Accordingly, these embodiments provide the technical effect of exhibiting or predicting turbine behavior at operating points or operating conditions that differ from current turbine operation. In addition, additional technical effects are provided that allow automated turbine control based at least in part on the modeled behavior and operating characteristics, which optionally turbines in these operator-specific conditions. It may include generating operator-specific scenarios, inputs, operating points, and / or operating conditions to predict behavior and operating characteristics. A further technical effect realized is the ability to predict various virtual scenarios that allow the operator to make more informed control and action decisions such as scheduling, loading, turndown, etc. including. As will be appreciated, reference is made herein to system step diagrams, methods, apparatus, and computer program products according to exemplary embodiments of the invention.

  Referring now to FIG. 17, a flow diagram 500 according to an alternative embodiment of the present invention is illustrated. As will be appreciated, flow diagram 500 includes aspects that may be used as part of a control method or control system to facilitate optimization of power plant 501. The power plant 501 can be similar to any of those discussed in connection with FIGS. 2 and 3, but unless otherwise limited in the appended claims, the invention is otherwise It should be appreciated that it could also be used in connection with a type of power plant. In a preferred embodiment, power plant 501 may include a plurality of thermal power generation units that generate electricity for sale within the power system market, such as those discussed in connection with FIG. Is. Power plant 501 can include many possible types of operating modes, for example, different schemes in which the plant's thermal power units are involved or operated, plant power levels, plant Includes methods that respond to changing ambient conditions while satisfying load requirements. It will be appreciated that the operating mode may be described and defined by operating parameters that take into account the physical characteristics of the particular aspect of operation of the power plant 501. As further illustrated in FIG. 17, the present invention can include a power plant model 502. The power plant model 502 can include a computerized representation of a power plant that correlates process inputs and outputs as part of a simulation intended to mimic the behavior of the plant. As shown, the present invention further includes a tuning module 503, a plant controller 505, a tuned power plant model 507, a plant operator module 509, and an optimizer 510, each of which: Will be discussed individually.

  Power plant 501 can include a sensor 511 that measures operating parameters. These sensors 511, as well as the operating parameters that these sensors 511 measure, can include any of those previously discussed herein. As part of the method, the sensor 511 may measure the operating parameters during the initial operational period, the current operational period, or the first operational period (hereinafter "first operational period"). , And those measurements can be used to tune the mathematical model of the power plant, which then models the subsequent operation, as discussed below. Of the optimization process for controlling the power plant 501 in a modified or optimized manner of operation during the first period of time or during the second period of operation (hereinafter “second period of operation”). It can be used as part. The measured operating parameters may themselves be used to evaluate plant performance, or may be used in calculations to derive performance indicators relating to particular aspects of power plant operation and performance. . As will be appreciated, this type of performance indicator can include heat rate, efficiency, power generation capacity, as well as others. Thus, as an initial step, the operating parameters measured by the sensor 511 during the first operating period may be used as one or more performance indicators (or calculate values for one or more performance indicators). Can be used to). As used herein, such values for performance indicators (ie, values based on measured operating parameters) will be referred to herein as "measured values." Measurements of operating parameters and / or measurements of performance indicators may be communicated 512 to both plant controller 505 and tuning module 503, as shown. Tuning module 503, from the data reconciliation or tuning process, for use in tuning power plant model 502 to construct tuned power plant model 507, as discussed in more detail below. Can be configured to calculate the feedback of the.

  Power plant model 502 can be a computerized model configured to simulate the operation of power plant 501, as discussed. According to the method, the power plant model 502 can be configured to simulate power plant operation corresponding to a first operating period of the power plant 501. To accomplish this, the power plant model 502 may be provided with information and data regarding operating parameters for the first operating period. This information may include any of the operating parameters measured during the first operating period, but the input data for power plant model 502 is limited to a subset of the measured operating parameters. It will be recognized that you will get. In this way, the power plant model 502 can then be used to calculate values for selected operating parameters that are excluded from the input dataset. More specifically, the power plant model may be provided with input data for the simulation, from which the input data for the simulation, including many of the measured values for the operating parameters, is selected. Certain measured values for operating parameters have been omitted. As an output, the simulation can be configured to predict simulated values for selected operating parameters. The method can then use the simulated values to predict values for performance indicators. In this case, these values for the performance index will be referred to herein as “predicted values”. In this way, the measurement value for the performance index determined directly from the measured power plant operating parameter can have a corresponding predicted value. As shown, the predicted value for the performance index may be communicated 514 to the tuning module 503.

  The tuning module 503 may be configured to compare the corresponding measured and predicted values for the performance index, and is adapted to determine the difference therebetween. As will be appreciated, the difference, when calculated accordingly, reflects the error level between the actual performance (or its measurement) and the performance simulated by the power plant model. Power plant model 502 may be tuned based on this difference or feedback 515. In this way, the tuned power plant model 507 is configured. The tuned power plant model 507 may also be referred to as an off-line model or a predictive model, and the tuned power plant model 507 may then perform subsequent operations by simulating proposed or possible modes of operation. Can be used to determine an optimized mode of operation for a period of time. Simulations can include estimates or expectations for future unknown operating conditions, such as ambient conditions. As will be appreciated, the optimization may be based on one or more performance goals 516, within which one or more performance goals 516 define a cost function. As shown, performance goals 516 may be communicated to optimizer 510 through plant operator module 509.

  The process of tuning the plant model can be configured as an iterative process that includes several steps. As will be appreciated, in accordance with certain embodiments, the power plant model 502 may be configured such that logic statements and / or parameterized equations provide process input (ie, fuel supply, air supply, etc.) process output. It is possible to include algorithms that correlate (electricity generated, plant efficiency, etc.). Tuning the power plant model 502 includes adjusting one of the algorithms in the power plant model 502 and then using the adjusted power plant model 502 to determine the impact of the adjustment. As such, simulating the operation of the power plant 501 for a first period of operation may be included. More specifically, the predicted value for the performance index may be recalculated to determine the effect that adjustments to the power plant model had on the calculated difference. If the adjusted power plant model 502 is used to find the difference is small, the power plant model 502 may be updated or "tuned" to include that adjustment in the forward direction. It will be further appreciated that the power plant model 502 may be constructed with multiple logical statements that include performance multipliers used to reflect changes in the manner in which the power plant operates under certain conditions. . In such cases, tuning the power plant model 502 based on the calculated differences includes a) making adjustments to one or more of the performance multipliers and b) adjusted performance multipliers. Simulating the operation of the power plant for a first operating period using a power plant model 502 having c, and c) a performance multiplier so as to determine whether the recalculation results in a reduction of the difference. Using the power plant model 502 as adjusted by the above, recalculating predicted values for performance indicators. These steps can be repeated until the adjustment made to one of the performance multipliers results in a reduction in the difference, which reduces the difference in that the model is Will be more accurately simulated. It will be appreciated that the performance multiplier may be related to, for example, expected performance degradation based on accumulated time of operation of the plant. In another example where the performance index includes power generation capacity, the step of tuning the power plant model 502 recommends adjustments to the coefficients based on the difference between the measured power generation capacity and the predicted power generation capacity. Can be included. Such adjustments can include changes that ultimately result in the predicted power generation capacity being substantially equal to the measured power generation capacity. Therefore, the step of tuning the power plant model 502 is such that the predicted or simulated value for the performance index is substantially equal to (or within the margin of the measurement for the performance index). Up to)) can include modifying one or more correlations in the power plant model 502.

  Once tuned, the method can then use the tuned model 507 to simulate the operation of the proposed power plant. According to a particular embodiment, the next step of the method involves determining which simulated behavior is preferable given the defined performance goals 516. In this way, the optimized mode of operating the power plant can be determined. According to a preferred embodiment, the process of determining the optimized mode of operation can include several steps. First, the plurality of proposed modes of operation may be selected or chosen from many possible ones. For each of the proposed operating modes, a corresponding proposed parameter set 517 may be generated for the second operating period. As used herein, a parameter set defines values for multiple operating parameters, and collectively, the parameter set is such that it defines or describes aspects of particular modes of operation. ing. As such, the proposed parameter set may describe many of the potential operating modes of the power plant 501, or may be configured to relate to many of the potential operating modes of the power plant 501, It can also be configured as an input data set for a tuned power plant model 507 to simulate operation. Once the operating parameters have been generated and organized into a proposed parameter set, the tuned power plant model 507 can simulate the operation of the power plant 501 according to each of the proposed parameter sets. Is. The optimizer 510 can then evaluate the results of the simulated action 519 for each of the proposed parameter sets 517. It is possible to carry out the evaluation according to the performance objective defined by the plant operator and the cost function defined therein. The optimization process can include any of the methods described herein.

  The cost function defined by the performance objective can be used to assess the economic performance of the simulated operation of the power plant 501 over the second operating period. Based on the evaluation, one of the proposed parameter sets may be regarded as producing a simulated behavior that is preferential as compared to that produced by the other proposed parameter sets. In accordance with the present invention, the mode of operation that corresponds to or is described by the proposed set of parameters that produces the most preferred simulated operation is designated as the optimized operation mode. Once determined, the optimized mode of operation may be passed to the plant operator for review, or communicated to the plant controller for automated implementation, as discussed further below.

  According to a preferred embodiment, the method of the present invention can be used to evaluate a particular mode of operation and determine and recommend preferred alternatives. As will be appreciated, the power generation unit of power plant 501 is controlled by an actuator having a variable set point, which is controllably linked to a control system such as plant controller 505. The operating parameters of the power plant 501 can be classified into three categories: manipulated variables, disturbance variables, and controlled variables. The manipulated variable evaluates a controllable process input that can be manipulated via an actuator to control the controlled variable, while the disturbance variable is an uncontrollable process input that affects the controlled variable. Evaluate. The controlled variable is the controlled process output for a defined target level. According to a preferred embodiment, the control method comprises receiving an expected value for the disturbance variable for the second period of operation (ie the period of operation for which the optimized mode of operation is being calculated). It is possible. Disturbance variables can include ambient conditions such as ambient temperature, pressure, and humidity. In such a case, the proposed parameter set generated for the second operating period may include values for the disturbance variable that are related to expected values for the disturbance variable. More specifically, the generated values for each ambient condition parameter can include a range of values for each ambient condition parameter. This range can include, for example, low, medium, and high cases. It will be appreciated that having multiple cases may allow a plant operator to plan for best / worst case scenarios. The expected values can include likelihood ratings corresponding to different cases, which allows plant operators to plan for various operational contingencies and / or hedge losses. It is possible to further support doing.

  The step of generating the proposed parameter set can include generating a target level for the controlled variable. Target levels can be generated to correspond to competing or alternative modes of operation of power plant 501 and can include operator input. Such operator input may be facilitated by the plant operator module 509. According to a preferred embodiment, such target level may include a desired power level for the power plant 501, which may occur subject to past usage patterns for the plant. It can be based on power level. As used herein, "power level" reflects the load level or level of electricity generated by the power plant 501 for commercial power distribution during the second period of operation. The step of generating the proposed parameter set may include generating a plurality of cases in which the power level remains the same or constant. Such a constant power level can reflect the base load for the plant or set of power generation units. Multiple target levels may be generated if each target level corresponds to a different level of involvement from each of the power generation units, and these are derived to obtain a possible mode of operation given historical usage. obtain. The method can then determine the most efficient mode of operation given known constraints. Additionally, the proposed parameter set can be generated such that the disturbance variable maintains a constant level for the multiple cases generated for each target level. The constant level for the disturbance variable can be based on the expected value received. In such a case, according to one aspect of the present invention, the step of generating the proposed parameter set is such that the optimal ambient conditions for achieving the base load level given the expected or expected ambient conditions. Generating a plurality of cases in which the manipulated variable is varied over each range to determine the optimized operating mode. According to an exemplary embodiment, the cost function may be defined as plant efficiency or heat rate, or may include more direct economic indicators such as operating costs, income, or profits. . Thus, the most efficient way to control the power plant 501 can be determined in situations where the base load is known and the disturbance variable can be predicted with a relatively high level of accuracy. In such a case, the optimized operating mode determined by the present invention is a particular control solution (ie, a particular setpoint, and / or , And therefore the range with respect to the actuators that control the manipulated variables of the power plant). When calculated in this way, the control solution represents an optimized mode of operation for satisfying the defined or contracted target load given the expected values for the various disturbance variables. This type of functionality can act as an optimization advisor or check for a period of days or between markets, which is a more efficient operation that still satisfies previously fixed load levels. Analyze the ongoing activity in the background for the purpose of finding the mode. For example, ambient conditions become known as the market period covered by previous power bids progresses, or at least the level of confidence that accurately predicts the ambient conditions is estimated during the bidding process. Increase beyond what you have. Given this, the method can be used to optimize the control solution to meet the load being fed, given a more solid knowledge of the ambient conditions. This particular functionality is illustrated in FIG. 17 as a second set of parameters 517 and as a simulated operation 519 associated with the second set of parameters 517. As such, the optimization process of the present invention may also include a "fine-tune" aspect whereby a simulation run on the tuned power plant model 507 is for a more efficient control solution. Advise, then it can be communicated to and implemented by the plant controller.

  Another aspect of the invention includes its use for optimizing fuel purchases for power plant 501. It will be appreciated that power plants typically make regular fuel purchases from fuel markets that operate in a particular manner. Specifically, such fuel markets are typically operated on a probabilistic basis, in which the power plant 501 predicts the amount of fuel needed for future periods of operation and then predicts. Make a purchase based on. In such a system, power plant 501 seeks to maximize profit by maintaining a low fuel inventory. However, the power plant 501 regularly purchases an excess amount of fuel and has an insufficient supply of purchased fuel to generate the amount of power the plant has contracted to provide during the power supply process. It is designed to avoid costly situations. This type of situation can occur, for example, when changing ambient conditions result in less efficient power generation than expected, or when the true power generation capacity of a power plant is overestimated. There is. Some aspects of the present application, which have already been discussed, determine an optimized mode of operation and use it to calculate a very accurate forecast for the required fueling. It will be appreciated that it can be used. That is, the present optimization process can provide a more accurate prediction of plant efficiency and load capacity, which can be used to estimate the amount of fuel needed for future operating periods. This allows the plant operator to maintain a smaller margin for fuel purchases, which benefits the economic performance of the plant.

  According to an alternative embodiment, the present invention comprises a method for optimizing plant performance, in which a forecast planning horizon is defined and used in the optimization process. As will be appreciated, the forecast planned horizon is a future period of operation, which for the purpose of determining the mode of operation optimized with respect to the initial time interval of the forecast planned horizon. It is divided into intervals that repeat on a regular basis. Specifically, the operation of the power plant is optimized by optimizing performance over the forecasted planning horizon, which in turn determines the mode of operation optimized for the initial time interval. used. As will be appreciated, the process then repeats to determine how the power plant should operate during the next time interval, which is to be recognized. As such, there is an initial time interval for the next iteration of the optimization cycle. For this subsequent optimization, the forecast planned horizon can remain the same, but is redefined relative to what is currently defined as the initial time interval. This means that the forecasted horizon can be effectively pushed forward with additional time intervals for each iteration. As already mentioned, the “proposal set of parameters” comprises values for a plurality of operating parameters, thereby defining or describing one of the possible operating modes for the power plant 501, Represents a dataset. According to a preferred embodiment, the process of determining an optimized mode of operation when participating in a forecasted horizon can include one or more of the following steps. First, a plurality of proposed planned horizon parameter sets are generated for the projected planned horizon. As used herein, a "proposal planned horizon parameter set" includes a proposed parameter set for each of the time intervals of the forecasted plan horizon. For example, a 24-hour forecast planned horizon can be defined as including 24 one-hour time intervals, where the proposed planned horizon parameter set is proposed for each of the 24 time intervals. It is meant to include a parameter set. As a next step, the proposed plan horizon parameter set is used to simulate the behavior over the forecast plan horizon. Then, for each of the simulation runs, the cost function is used to evaluate the economic performance, and which of the proposed planned horizon parameter sets is the most suitable, i.e., used herein. As described above, it is decided whether to represent the “optimized planning period simulation run”. According to an exemplary embodiment, the operating mode described in the optimized planning horizon simulation run for the initial time interval of the forecasted planning horizon is then set to the period of operation corresponding to the initial time interval. Can be designated as an optimized mode of operation for. The optimization process may then be repeated for subsequent time intervals. The invention may include receiving expected values for the disturbance variable for each of the time intervals defined within the forecasted horizon. Then, the proposed planned horizon parameter set is generated such that the proposed parameter set corresponding to each of the time intervals includes a value for the disturbance variable that is related to the expected value received for the disturbance variable. Can be done.

  As will be appreciated, the proposed planned horizon parameter set may be generated to cover a range of values for the disturbance variable. As mentioned above, the range can include multiple cases for each of the disturbance variables, and can include high and low values that represent above and below the expected value, respectively. According to any of the described embodiments, the steps of simulating a mode of operation and determining an optimized mode of operation from the simulation can be repeated and configured into an iterative process. Will be recognized. As used herein, each iteration is referred to as an "optimization cycle". It will be appreciated that each iteration may include defining a period of subsequent or next operation for optimization. This subsequent period may occur shortly after the period of operation optimized by the previous cycle, or for the purpose of preparing a case, for example a power supply bid, or an alternative maintenance schedule. For the purpose of advising on the financial impact of the, it is possible to include periods of action corresponding to future periods, as is the case when the method is used.

  The steps of tuning the power plant model 502 may be repeated to update the tuned power plant model 507. Thus, a tuned power plant model 507 that reflects recent tuning can be used with the optimization cycle to produce more effective results. According to alternative embodiments, the optimization cycle and the cycle for tuning the power plant model 502 may be isolated relative to each other, with each cycle repeating according to its own schedule. In other embodiments, the power plant model 502 may be updated or tuned after a predetermined number of optimization cycle iterations. The updated and tuned power plant model 507 is then used in subsequent optimization cycles until a predetermined number of iterations have taken place to initiate another tuning cycle. In a particular embodiment, the tuning cycle is performed after each optimization cycle. According to an alternative embodiment, the number of optimization cycles that initiate tuning of the power plant model 502 is related to the number of time intervals for the forecasted time horizon.

  The present invention, as stated, is able to optimize the operation of the power plant 501 according to performance goals, which may be defined by the plant operator. According to a preferred embodiment, the method is used to economically optimize the operation of a power plant. In such cases, performance goals include and define a cost function that provides a measure for economic optimization. According to an exemplary embodiment, the simulated behavior for each of the proposed parameter sets comprises, as output, the predicted value for the selected performance index. The cost function can include an algorithm that correlates the predicted value for the performance index with the operating cost or some other indication of economic performance. Other performance indicators that may be used in this manner include, for example, power plant heat rate and / or fuel consumption. According to an alternative embodiment, the simulation output includes predicted values for the hot gas path temperature for one or more of the thermal power generation units of power plant 501, which are the components consumed. It can be used to calculate the life cost. This cost reflects the predicted degradation cost associated with the hot gas path components resulting from the simulated operation. The cost function can further include an algorithm that correlates the predicted value for the performance index with the operating revenue. In such a case, operating revenues may then be compared to operating costs to reflect net revenues or net income for power plant 501. The method may further include the step of receiving an expected price for electricity sold in the market over the period of time being optimized, the selected performance indicator being the power level of electricity. The output level of electricity can then be included and can then be used to calculate the expected operating revenue for the period of the upcoming operation. Thus, the method can be used to maximize economic benefits by comparing operating costs and revenues.

  As will be appreciated, performance goals may be further defined to include selected maneuverability constraints. According to certain alternative embodiments, the method disqualifies any of the proposed parameter sets that produce a simulated behavior that violates any one of the defined usability constraints. Including the step of performing. Operability constraints include, for example, emission thresholds, maximum operating temperatures, maximum mechanical stress levels, etc., as well as legal or environmental regulations, contractual conditions, safety regulations, and / or machine or component operability thresholds and limits. Can be included.

  The method comprises, as already mentioned, generating a proposed set of parameters 517 describing alternative or possible modes of operation of the power plant 501. As shown, the proposed parameter set 517 can be generated in the plant operator module 509 and can include inputs from the plant manager or a human operator. Roughly speaking, the possible modes of operation can be considered as competing modes, for which the simulations are performed to determine the mode of operation that best meets the performance goals and assumed conditions. It has become. According to the exemplary embodiment, these alternative modes of operation may be selected or defined in several ways. According to the preferred embodiment, alternative modes of operation include various levels of power output for power plant 501. The power level, as used herein, refers to the level of electricity generated by the power plant 501 for commercial distribution in the market during a defined market period. The proposed parameter set can be configured to define multiple cases at each of the different power levels. Several power levels may be covered by the proposed parameter set and the power level chosen may be configured to match the range of potential power outputs for power plant 501. It will be appreciated that the range of possible output levels need not be linear. Specifically, due to the multiple power generation units of the power plant and their associated scalability limitations, the proposed parameter set is more feasible, given the particular configuration of the power plant 501. At the preferred level they can be grouped or focused.

  As mentioned, each of the competing modes of operation can include multiple cases. For example, if competing modes of operation are defined differently, multiple cases may be chosen, with the output level reflecting the different modalities achieved thereby. When a power plant has multiple power generation units, multiple cases at each power level may be differentiated by how each of the thermal power generation units is operated and / or involved. . According to one embodiment, some generated cases are differentiated by varying the percentage of the output level provided by each of the power generation units. For example, the power plant 501 can include a combined cycle power plant 501, in which the thermal power generation unit includes a gas turbine and a steam turbine. Additionally, gas turbines and steam turbines may be augmented with intake conditioning systems such as chillers and HRSG duct firing systems, respectively. As will be appreciated, the intake conditioning system may be configured to, for example, cool the inlet air of a gas turbine to enhance its power generation capacity, and the HRSG duct firing system may include: It can be configured as a secondary heat source for the boiler to enhance the power generation capacity of the steam turbine. According to this example, the thermal power generation unit comprises a gas turbine or, alternatively, a gas turbine augmented by an intake conditioning system and a steam turbine, or alternatively, steam augmented by an HRSG duct firing system. Including a turbine. Then multiple cases covered by the proposed parameter set include cases where these particular thermal power units are involved in different ways while still satisfying different power levels chosen as competing modes of operation. It is possible. The simulated behavior can then be analyzed according to defined criteria to determine which reflects the optimized mode of operation.

  According to an alternative embodiment, the proposed parameter set can be derived to obtain different operating modes and calculate the economic benefits of maintenance operations. To achieve this, one of the competing modes of operation may be defined as one in which the maintenance operation is assumed to be completed before the period of operation selected for optimization. . This mode of operation may be defined to reflect the performance gains that are expected to accompany the completion of this maintenance operation. An alternative mode of operation may be defined as one in which maintenance operations are not performed, which means that simulations of multiple cases for this mode of operation will not include expected performance enhancements. The results from the simulation can then be analyzed so that the economic effects are better understood, and multiple cases are used, with different scenarios (eg fuel price fluctuations or unexpected ambient conditions). , Etc.) can influence how they affect the results. As will be appreciated, using the same principle, competing modes of operation can include turndown and shutdown modes.

  The present invention further includes different schemes in which the optimization process may be used by a power plant operator to automate the process and improve efficiency and performance. According to one embodiment, as illustrated in FIG. 17, the method is calculated for human operator approval before the power plant 501 is controlled according to the optimized mode of operation. Communicating the optimized mode of operation 521 to the plant operator module 509. In advisor mode, the method may be configured to present alternative modes of operation and the economic ripple effects associated with each mode, making such alternatives noticeable to plant operators. There is. Alternatively, the control system of the method can function to automatically implement the optimized solution. In such cases, the optimized mode of operation may be electronically communicated to the plant controller 505 to facilitate control of the power plant 501 in a corresponding manner. In a power system that includes an economical power supply system for distributing power generation between groups of power plants 501, the optimization method of the present invention is more accurate and competitive for submission to a central authority or power dispatching office. It can be used to generate bids. As will be appreciated by those skilled in the art, the previously described optimization schemes generate bids that reflect true power generation capacity, efficiency and heat rate while at the same time the power plant chooses different modes of operation. It can be used to provide useful information to the plant operator regarding economic trade-offs made in future market periods. This type of precision improvement and additional analysis minimizes the risk that a power plant will remain competitive in the bidding process while at the same time obtaining unprofitable power supply results due to unexpected contingencies. Help to ensure.

  18-21 illustrate exemplary embodiments of the present invention relating to power plant turndown and / or shutdown operations. As illustrated in the flow diagram 600 of FIG. 18, which may be referred to as a “turndown advisor,” the first embodiment provides for a defined or selected period of operation (“selected operating period”). It teaches methods and systems for simulating and optimizing turndown levels for power plants in between. In a preferred embodiment, the method is used with a power plant having multiple gas turbines, which may include a combined cycle plant having multiple gas turbines and one or more steam turbines. The tuned power plant model can be used to determine an optimized minimum load for operating the power plant at turndown levels during selected operating periods. As mentioned above, an "optimized" operating mode may be defined as an operating mode that is considered to be preferred or evaluated over one or more other possible operating modes. Modes of operation for the purposes of these embodiments include allocation of specific power generation units to meet load start-up or shutdown plans or other performance goals, as well as the physical configuration of the power generation units within the power plant. Is possible. Such functionality allows the present invention to consider a large number of plant combinations when reaching an optimized or enhanced mode of operation, where a large number of plant combinations are provided for each power generation unit. Means to allow one or more of the units to shut down while leaving other units operating at full or turndown levels. The method further enforces other constraints, such as maneuverability constraints, performance goals, cost functions, operator inputs, and ambient conditions in the calculation of enhanced turndown operating modes for power plants that enhance performance and / or efficiency. It can be taken into consideration. The method is for the optimization of turndown operating modes, as described herein and / or exactly as set forth in the appended claims, for current ambient conditions. And predicted ambient conditions can be taken into account, and one of the power generation units can be changed when the unit configuration and / or control is changed and actual conditions deviate from those predicted conditions. It is adapted to dynamically adjust one or more movements. According to a preferred embodiment, such performance is at least partially defined as minimizing the level of fuel usage or consumption over the proposed turndown period of operation.

  The turndown advisor of the present invention takes into account several factors, criteria, and / or operating parameters in reaching an optimized or enhanced turndown solution and / or a recommended turndown effect. It is possible. According to a preferred embodiment, these include but are not limited to gas turbine engine operating limits (ie temperature operating limits, aerodynamic operating limits, fuel split operating limits, lean blowout operating limits, mechanical Operating limits and operating limits of emission limits), gas turbine and steam turbine control systems, minimum steam turbine throttle temperatures, maintenance of vacuum seals on condensers, and configurations or lineups of systems or system controls. Including other factors such as: One of the outputs of the optimization may include a recommended operating mode and configuration of one or more power plants, which may be wind, solar, reciprocating engines, nuclear. , And / or other types, including different types of power plants. It will be appreciated that the recommended mode of operation may be automatically initiated or electronically communicated to the plant operator for approval. Such control may be implemented via an off-premise or on-premise control system configured to control the operation of the power generation unit. Additionally, in situations where the power plant includes multiple gas turbine engines, the output of the method should be such that during the turndown period, which of the gas turbines should continue to operate and which should be shut down. Can be included, which is the process discussed in more detail in connection with FIG. For each gas turbine for which the advisor recommends continued operation during the turndown period, the method can further calculate the load level. Another output may include calculating the total load for the power plant during the turndown period, and, as stated, adjusted if conditions change and is expected. It may include calculating an hourly target load profile based on the ambient conditions. The invention is also able to calculate the predicted fuel consumption and emissions of a power plant during a turndown period. The output of the disclosed method can include an operating lineup / configuration subject to the control set points available to the power generation unit and plant, resulting in more efficient realization of the target power generation level. ing.

  As discussed above, traders and / or plant managers (hereinafter, unless otherwise differentiated, "plant operators") are not bound by existing contract terms, typically such as electricity spot markets. Bid those power plants in a promising market. As an additional consideration, the plant operator is responsible for ensuring that sufficient fuel supply is maintained so that the power plant can meet the target or contracted power generation levels. However, in many cases, the fuel market behaves in the hope that favorable price terms will be available to power plants that are willing or able to promise future fuel purchases in advance. Has become. More specifically, the sooner the fuel is purchased, the more favorable the price. Given these market dynamics, in order for a power plant to achieve optimized or high levels of economic benefits, plant operators need to use other power generation units to utilize their power generation capacity. You have to bid the plant competitively while at the same time 1) so that the fuel can be pre-purchased to ensure a lower price, and 2) without the need for a large fuel buffer, Accurately estimate the fuel needed for future power generation periods so that a low fuel inventory can be maintained. If successful, plant operators will ensure better prices by early contracting future fuel purchases, while avoiding over-purchasing or fueling unnecessary and costly reserve fuels. There is no shortage of purchases that run the risk of shortage.

  The method of the present invention, in particular, by identifying the IHR profile for a particular configuration of a power generation unit or plant, as these relate to the preparation of power supply bids to secure a share of the power generation market, It is possible to optimize or enhance efficiency and profitability. The method can include identifying optimal power generation allocations across multiple power generation units in a power plant, or across several plants. The method is able to take into account the operational and control configurations available for those power generation units, swapping out possible arrangements, thereby reducing or minimizing if selected. It is possible to realize bidding that allows for the generation of electricity over the bidding period at an incurred cost. In doing so, the method can take into account all applicable physical, regulatory and / or contractual constraints. As part of this overall process, the method can be used to optimize or enhance turndown and shutdown operations for power plants with multiple power generation units. This procedure involves taking into account, for example, assumed extrinsic conditions such as weather or ambient conditions, gas quality, reliability of the power generation unit, and attendant obligations such as steam generation. It is possible. The method is used to enumerate IHR profiles for multiple generation units with multiple configurations, as well as control setpoints for selected turndown configurations, and then to identify possible extrinsic conditions in preparation for a plant feed bid. It is possible to control in order.

  One common decision for operators concerns whether to turn down or shut down a power plant during off-peak periods, such as at night, when demand or load requirements are minimal. As will be appreciated, the outcome of this decision depends largely on the plant operator's understanding of the economic spillover effects associated with each of these possible modes of operation. In certain cases, the decision to turn down the power plant can be readily apparent, but the optimal minimum load to maintain the power plant during the turn down period remains uncertain. That is, while the plant operator has made the decision to turn down the power plant over a specific period of time, the operator is confident about the turn down operating point that will run some of the power plant's power generation units in the most cost effective manner. I can't hold it.

  The turndown advisor of FIG. 18 may be used as part of a process to recommend an optimal minimum load to operate a power plant. This advisor function can further recommend best practices for a power plant given ambient conditions, economic inputs, and specific scenarios of operating parameters and constraints. From these inputs, the process can calculate the best operating level, as will be discussed in more detail in connection with FIG. 19, and then the operating parameters needed to control the power plant. Can be recommended. As will be appreciated, this functionality can result in some ancillary benefits, which include extended component life, more efficient turndown operation, and economics. Includes improved performance and improved accuracy when making fuel purchases.

  As illustrated in flow diagram 600, certain information and associated criteria may be gathered during the initial steps. At step 602, data, variables, and other factors associated with power plant systems and power generation units may be determined. These can include any of the factors or information listed above. According to a preferred embodiment, an ambient profile may be received, which may include an estimate of ambient conditions during the selected operating period. Also, relevant emissions data can be collected as part of this step, which can include emission limits for power plants as well as emissions to date. Another factor includes data related to potential sales of electricity and / or steam during selected operating periods. Other variables that may be determined as part of this step include the number of gas turbines in the plant, the number of combustion and control systems for each of the gas turbines, and any other that may be relevant to the calculations discussed below. Including plant-specific restrictions in.

  At step 604, the duration of the proposed turndown operation (or "selected operation duration") may be defined in detail. As will be appreciated, this can include an operating period defined and selected by the user or plant operation, during which the available turndown operating modes are available. Analysis is desired. The definition of the selected operating period is its expected length, as well as a user-specified start time (ie, the time that the selected operating period will start) and / or stop time (ie, selected). The time period during which the operation period ends) can be included. This step may further include defining an interval within the selected operating period. The intervals may be configured to subdivide the selected operational period into a plurality of sequential, regularly spaced time periods. For the examples provided herein, the interval will be defined as one hour and the selected operating period will be defined as including a plurality of one hour intervals.

  At step 606, the number of gas turbines involved in the optimization process for the selected operating period may be selected. This may include all or some parts of the gas turbine in the power plant. As described in more detail below, the method further includes consideration of other power generation units in the power plant, such as steam turbine systems, etc., taking into account their operating conditions during the selected operating period. It is possible to put in. Determining the gas turbine involved in the turndown operation may include prompting or receiving input from a plant operator.

  In step 608, the method may construct a permutation matrix given the number of gas turbines determined as part of the proposed turndown operation during the selected operation period. As will be appreciated, the permutation matrix is a matrix that includes various manners in which multiple gas turbine engines may be engaged or operated during a selected operating period. For example, as illustrated in the exemplary permutation matrix 609 of FIG. 18, the permutation matrix for the two gas turbine cases includes four different combinations covering each of the possible configurations. Specifically, if the power plant includes a first gas turbine and a second gas turbine, the permutation matrix includes the following rows or cases: a) first gas turbine and second gas turbine Both turbines are "on", ie are operated in a turndown operation; 2) both first gas turbine and second gas turbine are "off", ie shutdown operation 3) the first gas turbine is "on" and the second gas turbine is "off"; and 4) the first gas turbine is "off". , The second gas turbine is "on". It will be appreciated that in the case of a single gas turbine only two permutations are possible, while for three gas turbines seven different rows or cases are possible, which Each represents a different configuration of how the three gas turbine engines may be involved during a particular time frame in terms of "on" and "off" operating conditions. In the context of the optimization process discussed in connection with FIG. 17 and in the text associated with FIG. 17, each case or row of permutation matrices is considered to represent a different or competing mode of operation. obtain.

  As part of the steps represented by steps 610, 613, 614, 616, and 618, the method can configure the proposed parameter set for the proposed turndown operation. As mentioned, the selected operating period may be divided into several hourly time intervals. The process for constructing the proposed parameter set may begin at step 610, where it is determined whether each of the intervals has been addressed. If the answer to this question is yes, then the process may continue to the output step (ie, step 611), as shown, where the output of the turndown analysis is output. Are provided to the operator 612. If all of the intervals are not covered, then the process may continue to step 613, where one of the as yet uncovered intervals is selected. Then, in step 614, ambient conditions may be set for the selected interval based on the received expectations. Following step 616, the process may select a row from the permutation matrix and in step 618 the gas turbine on / off state may be set according to the particular row.

  From there, the method can continue along two different paths. Specifically, the method may continue with the optimization step represented by step 620, while also following the decision step in step 621, in which the process proceeds to all permutations of the permutation matrix. Or determine if the row was covered for the selected interval. If the answer to this is no, then the process may loop back to step 616, where a different replacement row for the interval is selected. If the answer to this is yes, then the process may continue to step 610, as shown, to determine if all of the intervals have been covered. There is. As will be appreciated, once all of the rows of the permutation matrix for each interval have been addressed, the process can proceed to the output step of step 611.

In step 620, the method can use a tuned power plant model to optimize performance, as previously discussed in FIG. Consistent with this approach, multiple cases may be generated for each of the competing modes of operation, that is, for each row of the permutation matrix for each of the intervals of the selected operating periods. According to a preferred embodiment, the method produces a proposed parameter set, in which some operating parameters are varied so as to determine their influence on the selected operating parameter or performance index. For example, according to this embodiment, the proposed parameter set may include manipulating set points for inlet guide vanes (“IGV”) and / or turbine exhaust temperature (“T _exh ”). Yes, to determine which combination produces a minimized total fuel consumption rate for a power plant given the ambient condition expectations for a particular row on / off state and a particular interval. . As will be appreciated, the motion that minimizes fuel consumption while satisfying other constraints associated with turndown motion represents one modality, which modality results in one turndown performance. Or it may be economically optimized, or at least economically enhanced, for multiple alternative modes of operation.

As shown, according to particular embodiments, cost functions, performance goals, and / or manipulative constraints may be used by the present invention during this optimization process. These may be provided via the plant operator and are represented by step 622. These constraints can include limits on IGV setpoints, T _exh limits, combustion limits, etc., as well as limits associated with other thermal systems that may be part of the power plant. For example, in a power plant having a combined cycle system, steam turbine operation or maintenance during turndown operation may present certain constraints, such as, for example, maintaining minimum steam temperature or condenser vacuum seals. is there. Another operability constraint is that certain ancillary systems may be affected in certain operating modes, and / or certain subsystems such as evaporative coolers and chillers are mutually exclusive. Yes, it can include the necessary logic.

  Given the spacing of the permutation matrix and the different rows, as the method cycles through the iterations, the results of the optimization can be communicated to the plant operator in step 611. These results can include optimized cases for each of the rows of the permutation matrix for each of the time intervals. According to one example, the output describes optimized behavior, which is defined by a cost function of fuel consumption for the power plant for each of the permutations for each of the intervals. Specifically, the output is a possible plant configuration (as represented by the rows of the permutation matrix) for each interval, while also satisfying the manipulative constraints, performance goals, and assumed ambient conditions. It is possible to include each of the required minimum fuels (optimized using a tuned power plant model according to previously described methods). According to another embodiment, the output includes optimizations that, in the same manner, minimize the power output level (ie, megawatts) for the potential plant configuration for each of the intervals. As will be appreciated, some of the possible plant configurations (as represented by the permutation of the permutation matrix) impose manipulability constraints regardless of the fueling to generate the power level. May not be satisfied. Such results may be discarded, not considered further, or reported as part of the output of step 611.

  FIGS. 19 and 20 show that the gas turbine of a power plant may be operated over a selected operating period that includes a defined interval (“I” in the figure), subject to typical constraints associated with transient operation. The scheme is represented diagrammatically. As will be appreciated, transient operation includes switching the power generation unit between different modes of operation, including with or out of shutdown mode of operation. As shown, multiple operating paths or sequences 639 may be implemented depending on: 1) the initial state 640 of the gas turbine; and 2) at intervals that may change subject to transient operating constraints. A decision made about whether to change the operating mode. As will be appreciated, a number of different sequences 639 represent multiple ways in which the power generation unit may be operated over the intervals shown.

  As will be appreciated, the output of the method of FIG. 18 may be used in conjunction with the diagrams of FIGS. 19 and 20 to configure the proposed turndown operating sequence for a power generation unit of a power plant. It is possible. That is, FIGS. 19 and 20 show examples of how a power generation unit of a power plant may be involved and how its operating mode may be modified over time. It is shown that it can be used if the operating mode of the power generation unit remains unchanged, if the operating mode of the unit is modified from the shutdown operating mode to the turndown operating mode, and if the operating mode of the unit is from the shutdown operating mode. It is possible to include a case where the turndown operation mode is modified. As shown, the transient motion constraint used in this example requires that the mode of operation be modified so that the unit stays in the modified mode of operation for at least two of the intervals. That's what it means. Many sequences (or paths) for a power generation unit to reach the last interval represent possible turndown sequence of operations available to the unit subject to transient operation constraints.

  As will be appreciated, the analysis result from FIG. 18, ie, optimized turndown behavior for each of the matrix permutations, selects multiple preferred cases from the possible turndown behavior sequences. Can be used to do so, which can be referred to as the proposed turndown operation sequence. Specifically, given the results of the method described in connection with FIG. 18, the proposed turndown operation sequence is performed according to a selected cost function (eg, MW power or fuel consumption, etc.). Can also be selected from the case of turndown behavior that satisfies the plant performance goals and constraints. The considerations illustrated in FIGS. 19 and 20 represent a scheme for determining whether a turndown motion sequence is achievable subject to transient motion constraints. That is, the proposed turndown operating sequence reached by the combinatorial analysis of FIGS. 18-20 is an operating sequence that meets the time constraints associated with moving a unit from one operating mode to another. is there.

  Referring now to FIG. 21, a method is provided for further modeling and analyzing turndown behavior of a power plant. As will be appreciated, this method can be used to analyze turndown vs. shutdown costs for a particular case involving a single power generation unit over a defined time interval. However, it can also be used to analyze plant-level costs, where recommendations are needed for the manner in which the operation of some power generation units can be controlled over a selected period of operation with multiple intervals. Thus, the outputs of FIGS. 18 and 20 may be assembled to constitute a possible operating mode or sequence over spans of intervals, which, as will be demonstrated, then Twenty-one methods can be analyzed to provide a better understanding of turndown operation over a wider operational period.

  The plant operator must regularly determine between turndown and shutdown modes of operation during off-peak hours, as discussed above. Although certain conditions may make the decision straightforward, in many cases, among other things, the increased complexity of modern power plants and the multiple thermal power plants typically contained within each Given the unit, it is difficult. The decision to turn down or shut down a power plant, as will be appreciated, depends in large part on the full recognition of the economic benefits associated with each mode of operation. The present invention, in accordance with an alternative embodiment illustrated in FIG. 21, provides a plant with improved understanding of the trade-offs associated with each of these different modes of operation to enhance decision making. Can be used by the operator. According to certain embodiments, the method of FIG. 21 may be used in conjunction with the turndown advisor of FIG. 18 to enable a combined advisor function, which is 1) known in the art. Given the conditions and economic factors, the best bet between turndown and shutdown modes of operation is recommended for the power plant's power generating units and 2) turndown is the best bet for some of those units. If so, the optimum minimum turndown load level is recommended. In this way, the plant operator determines the units of a power plant based on which represents the economic best bet for the power plant, given the particular scenario of ambient conditions, economic inputs and operating parameters. It is possible to more easily identify situations in which one should turn down rather than shut down, or vice versa. There are also ancillary benefits such as extending the life of component parts. It should also be appreciated that the methods and systems described in connection with Figures 18 and 21 can be used separately.

  In general, the method of flow diagram 700 can be part of a "turndown advisor," or can be referred to herein as a "turndown advisor," and the method of flow diagram 700 can be , Applying data from user inputs and analytical operations to perform calculations to evaluate the costs associated with turning down the power plant and shutting down the power plant. It will be appreciated that the flow diagram 700 of FIG. 21 illustrates this advisory mechanism by utilizing the tuned power plant model discussed in detail above, according to certain preferred embodiments. provide. As part of this functionality, the present invention advises on various consequences between turning down and shutting down a power plant during off-peak demand periods, economic consequences and other consequences. It is possible. The present invention may provide relevant data that reveals whether it is preferable to turn down a power plant over shutting down the power plant over a particular market period. According to a particular embodiment, an operation with a lower cost may then be recommended to the plant operator as an appropriate measure, but as presented here, incidental issues that may influence the decision. Or other considerations may be communicated to the plant operator. The method can generate potential costs, as well as the probabilities of incurring such costs, and these considerations can influence the final decision as to which operating mode is preferred. There is a nature. Such considerations can include, for example, a complete analysis of both short-term operating costs as well as long-term operating costs associated with plant maintenance, operating efficiency, emission levels, equipment upgrades, and the like.

  As will be appreciated, the turndown advisor is implemented using many of the systems and methods described above, and in particular, the systems and methods discussed in connection with FIGS. 16-20. obtain. The turndown advisor of FIG. 21 can collect and use one or more of the following types of data: User-specified start and stop times for the proposed turndown period of operation (ie, , Turndown operating modes are analyzed or considered); Fuel costs; Ambient conditions; Time-off breakers; AC power use; Electricity or steam sales / price during the relevant period; Operation and maintenance over that period Costs; User Inputs; Calculated Turndown Loads; Expected Emissions for Operation; Current Emission Levels Expended by Power Plants and Limits for Defined Regulatory Periods; Specifications for Turning Gear Operation; Purge Process Related regulations and equipment; power plant operation Fixed costs for modes; costs associated with startup operations; plant startup reliability; imbalance charges or fines for delayed startups; emissions associated with startups; fuel used for auxiliary boilers when steam turbines are present Ratio; and historical data regarding how the power plant gas turbine was operating in the turndown and shutdown modes of operation. In certain embodiments, as discussed below, the output from the present invention can include: a recommended mode of operation (ie, turndown operation) for the power plant over the relevant time period. Modes and shutdown operating modes); costs associated with each operating mode; recommended plant operating load and load profile over time; recommended time to initiate unit start-up; and emissions consumed year-to-date And the remaining emissions for the rest of the year. According to a particular embodiment, the present invention is capable of calculating or predicting fuel consumption and emissions of a power plant over a relevant time period, which in turn for one or more particular gas turbine engines. Can be used to calculate the cost of turndown vs. shutdown. The method can use the cost of each gas turbine in shutdown mode and turndown mode to determine the combination with the lowest operating cost. Such optimization can be based on different criteria, which can be defined by the plant operator. For example, the criteria can be based on income, net income, emissions, efficiency, fuel consumption, etc. Additionally, according to an alternative embodiment, the method determines whether to take a purge credit; the gas turbine unit to be shut down and / or the gas turbine unit to be turned down (which may be, for example, It is possible to recommend certain measures, such as historical start-up reliability and potential imbalance charges that may be imposed due to delayed start). The present invention may be further used to enhance forecasts related to fuel consumption, making more accurate prospective fuel purchases, or, alternatively, fuel purchases for further future market periods. Which should have a positive impact on fuel prices and / or maintenance of lower fuel inventory or margins.

  FIG. 21 illustrates an exemplary embodiment of a turndown advisor according to an exemplary embodiment of the present invention, which is in the form of a flow diagram 700. The turndown advisor may be used to advise on the relative cost of shutting down the power plant or a portion thereof over future periods of operation while operating the power generation unit in turndown mode. According to this exemplary embodiment, the possible costs associated with shutdown and turndown modes of operation may be analyzed and then communicated to the plant operator for appropriate action.

  As an initial step, certain data or operating parameters may be gathered that may affect or be used to determine the operating cost during the selected turndown operating period. These are grouped accordingly between turndown data 701; shutdown data 702; and common data 703, as shown. Common data 703 includes those cost items related to both shutdown and turndown modes of operation. The common data 703 includes, for example, the selected operation period, and the turn-down operation mode is analyzed over the selected operation period. Two or more selected operating periods are defined with respect to competing modes of turndown operation and can be analyzed separately, so that a wider range of optimizations are achieved over an extended time frame. Will be recognized. As will be appreciated, defining the selected operating period may include defining the length of the period and its starting or ending point. Other common data 703 can include data about fuel prices, various emission limits for power plants, and ambient conditions, as shown. Regarding emission limits, the data collected will indicate that the limits that can be imposed during a defined regulation period, such as one year, the amount already generated by the power plant and the applicable regulation period already It can include the degree applied. Further, emissions data may include fines or other costs associated with exceeding any of the limits. As such, the method may be informed of the current status of the power plant for annual or periodic regulatory restrictions, as well as the potential for violations and the fines associated with such non-compliance. It is possible. This information may be relevant to a decision as to whether to shut down or turn down a power generation unit when each type of operation affects plant emissions differently. With respect to ambient condition data, such data can be obtained and used according to those processes already described herein.

  The turndown mode of operation has data uniquely associated with determining the operating cost associated with the turndown mode of operation, as will be appreciated. Such turndown data 701 includes income that may be earned through the power generated while the power plant is operating at the turned down level, as shown. More specifically, the turndown mode of operation is a mode of operation at which power generation continues, albeit at a lower level, so there is the potential for that power to generate revenue for the power plant. To the extent that this is done, the revenue can be used to offset some of the other operating costs associated with the turndown operating mode. Accordingly, the method includes receiving a price or other economic indication associated with the sale or commercial use of the electricity generated by the plant while operating in the turndown mode. This can be based on historical data and the income earned can depend on the turndown level at which the power plant operates.

  Turndown data 701 can further include operations and maintenance associated with operating a plant at a turndown level during a selected operating period. It can also be based on historical data, and such costs can depend on the turndown level for the power plant and how the power plant is configured. In some cases, this charge may be reflected as an hourly cost that depends on historical records of load levels and similar behavior. The turndown data 701 can further include data related to plant emissions while operating in turndown mode.

  The shutdown data 702 also includes some items specific to the shutdown mode of operation, and this type of data may be collected at this stage of the method. According to a particular embodiment, one of these is data relating to the operation of the turning gear during the shutdown period. Additionally, data regarding various aspects of shutdown operation will be defined. This can include, for example, data related to: the length of time required to bring the power generation unit from normal load levels to the engaged gears. The shutdown operation itself; which may include historical data; the length of time the power plant remains shut down according to the selected operating period; the length of time the power generation unit typically stays on the turning gear And data regarding the process in which the power generation unit is shut down and then restarted or brought back online, as well as the time required to do this, start-up fuel requirements, and start-up emissions data. In determining the start-up time, information such as the types of possible start-ups for the power generation unit and their associated specifications may be determined. As one of ordinary skill in the art will appreciate, the start-up process may depend on the time the power plant remains shut down. Another consideration that affects start-up time is whether the power plant includes certain features that can affect or reduce start-up time, and / or the operator of the power plant may The question is whether to choose to be involved in either. For example, the purging process can increase the start-up time if needed. However, purge credits may be available if the power plant is shut down in a particular manner. Fixed costs associated with shutdown operations include fixed costs associated with startup and may be identified during this step, along with costs specific to any of the associated power generation units. Emissions data associated with power plant startups and / or shutdowns may also be confirmed. These may or may not be based on historical records of operations. Finally, data relating to start-up reliability for each of the thermal power generation units can be confirmed. As will be appreciated, if the process of bringing the unit back online includes delays that result in the power plant failing to meet its load obligations, the power plant may charge fees, fines, and / or Damages may be imposed. These costs may be determined and may be considered in light of historical data related to startup reliability, as discussed in more detail below. As such, such fees are discounted to reflect the likelihood of burden and / or spending to hedge the risks of such fees or to insure against such risks. May include.

  From the initial data collection steps 701 to 703, the exemplary embodiment illustrated in FIG. 21 can proceed through a turndown analyzer 710 and a shutdown analyzer 719, each of which corresponds to May be configured to calculate an operating cost for the operating mode to be performed. As shown, each of these analyzers 710, 719 can proceed towards providing cost, emissions, and / or other data to step 730, where the potential Data regarding certain turndown and unit shutdown scenarios are compiled, compared, and finally output can be provided to the power plant operator in step 731. As will be discussed, this output 731 can include costs and other considerations for one or more of the possible scenarios, and ultimately the particular measures and It is possible to recommend the reason.

  With regard to turndown analyzer 710, the method can first determine the load level for the proposed turndown operation during the selected operation period. As discussed further below, most of the costs associated with turndown operation can be highly dependent on the load level at which the power plant operates, and how the plant will generate that load. Can be heavily dependent on how different thermal power units are involved (ie, which thermal power units are turned down and which are shut down), for example. Can be included. The turndown load level for the proposed turndown operation may be determined in a number of different ways according to alternative embodiments of the invention. First, the plant operator is able to select the turndown load level. Second, load levels can be selected through analysis of historical records regarding past turndown levels at which the plant operated efficiently. From these records, suggested load levels are available, for example, for efficiency, emissions, achievement of one or more site-specific goals, alternative commercial uses for power generated during turndown conditions. It can be analyzed and selected based on operator-supplied criteria such as gender, ambient conditions, and other factors.

  As a third method of selecting a turndown level for the proposed turndown operation, a computer-implemented optimizing program, such as the one described in connection with FIG. 18, may optimize the turndown level. Can be used to calculate In FIG. 21, this process is represented by steps 711 and 712. An optimized turndown level may be calculated by proposing a turndown mode of operation in step 711, and then in step 712 by analyzing whether operating limits for the power plant are met. . As will be appreciated, a more detailed description of how this is achieved is provided above in connection with FIG. By optimizing the turndown level using such a process, the turndown operating mode selected for comparison to the shutdown alternative mode for the selected operating period represents the optimized case. It will be appreciated that, and given this, the comparison between the turndown alternative mode and the shutdown alternative mode would be meaningful. As described in connection with FIG. 18, the minimum turndown level may be calculated via an optimization process that optimizes the turndown level according to operator-selected criteria and / or cost functions. One of the functions can be the level of fuel consumption during the proposed turndown operation period. That is, the optimized turndown level may be determined by optimizing fuel consumption towards a minimum level while also satisfying all other operating limits or site-specific performance goals.

  From there, the method of FIG. 21 determines the cost associated with the proposed turndown mode of operation for the selected period of operation according to the characteristics of the turndown mode of operation determined via steps 711 and 712. It is possible. As shown, step 713 can calculate the fuel consumption and then the fuel cost for the proposed turndown operation. According to a positively discussed exemplary embodiment that describes optimization based on minimizing fuel consumption, the fuel cost simply obtains the fuel level calculated as part of the optimization step. And then by multiplying by a price for the envisaged or known fuel. In the next step (step 714), the revenue derived from the electricity generated during the selected operating period will increase the proposed turndown level and commercial availability during the selected operating period. It can be calculated as a premise. Then, in step 716, operating and maintenance costs may be determined. The operating and maintenance costs associated with the proposed turndown operation may be calculated via any conventional method and may depend on the turndown level. Operating and maintenance costs can be reflected as hourly charges derived from historical records of turndown operations, and are part of the expected life of various component systems used during the proposed turndown operation. It is possible to include a component usage charge that reflects the. In the next step, indicated by step 717, the net cost for the proposed turndown mode of operation for the selected period of operation is the sum of costs (fuel, operation, and maintenance) and income. It can be calculated by subtracting.

  The method also includes step 718, which determines plant emissions over a selected operating period given the proposed turndown mode of operation, which is referred to as the "emissions impact". obtain. The net cost and emissions impact may then be provided to a compile and compare step, which is represented as step 730, so that the cost and emissions impact of different turndown scenarios may be analyzed. And finally recommendations can be provided at output step 731, as discussed further below.

  Turning to the shutdown analyzer 719, the shutdown analyzer 719 is for calculating aspects related to operating one or more of the power plant's power generation units in a shutdown mode of operation during a selected operating period. Can be used for. As part of this aspect of the invention, operations including shutting down a power plant and then restarting at the end of a selected time period may be analyzed for cost and emissions. According to a preferred embodiment, the shutdown analyzer 719 can determine the proposed shutdown mode of operation as part of the initial steps 720 and 721, and the proposed shutdown mode of operation is optimized. It is possible to represent the shutdown operation mode. The proposed shutdown mode of operation involves the process of shutting down one or more of the power generating units and then restarting the units to bring them back online at the end of the selected operating period. As will be appreciated, the length of time period during which the power generation unit is not operating can determine the type of start-up process that the power generation unit may have available. For example, whether hot or cold startup is available depends on whether the shutdown period is short or long, respectively. In determining the proposed shutdown operating mode, the method is capable of calculating the time required by the start-up process to bring the power generating unit back to the operating load level. At step 721, the method of the present invention can be checked to ensure that the proposed shutdown operating procedure meets all operating limits of the power plant. If one of the operating limits has not been met, the method can return to step 720 to calculate an alternative start-up procedure. This can be repeated until an optimized start-up procedure that satisfies the operating limits of the power plant has been calculated. As will be appreciated, in accordance with the methods and systems discussed above, a tuned power plant model can be used to simulate alternative shutdown modes of operation and associated operating periods and It is designed to determine the optimized case based on the project ambient conditions.

  Given the proposed shutdown mode of operation of steps 720 and 721, the process can continue by determining the costs associated with it. The initial steps include analyzing the characteristics of the startup process that the shutdown mode of operation includes. At step 722, the process can determine certain operating parameters of the start-up, which can include a determination as to whether a purge is required or required by the plant operator. Given the determined start-up, fuel costs may be determined in step 723. According to an exemplary embodiment, shutdown analyzer 719 then calculates a cost associated with the delay that may be incurred during the startup process. Specifically, as shown in step 724, the process may calculate the probability of such a delay. This calculation may include as inputs the type of start-up, as well as a historical record of past start-ups of the associated power generation units in the power plant, as well as data on the start-up of such power generation units in other power plants. . As part of this, the process can calculate the costs associated with the proposed shutdown mode of operation, which reflects the probability of a start delay and any fines such as the amount of damages that will be imposed. Is. This cost can include any costs associated with hedging tactics, which allow the power plant to pass on to service providers or other insurance companies a portion of the risk of incurring such fines.

  At step 726, the method may determine a cost associated with operating the turning gear during the shutdown process. Given the shutdown period, the method is able to calculate a speed profile for the turning gear, which is used to determine the cost for the auxiliary power needed to operate the turning gear. It As will be appreciated, this represents the power required to maintain the turning of the gas turbine rotor blades as they cool, which prevents warpage or deformation. Warping or deformation would otherwise occur if the blade were allowed to cool in the rest position. At step 727, operating and maintenance costs for the shutdown operation may be determined, as shown. The operating and maintenance costs associated with the proposed shutdown can be calculated via any conventional method. Operating and maintenance costs can include component usage fees that reflect a portion of the expected life of the various component systems used during a proposed shutdown operation. In the next step, indicated by step 728, the net cost for the proposed shutdown mode of operation for the selected period of operation is by adding the fuel, turning gear, and the determined costs of operation and maintenance. Can be calculated. The method may also include step 729, where step 729 determines the plant emissions over the selected operating period given the proposed shutdown mode of operation, which, as described above, It can be referred to as the “effect of emission amount” of the operation mode. The net cost and emissions impact may then be provided to the compilation and comparison step of step 730.

  At step 730, the method may compile and compare various plant turndown operating modes for the selected operating period. According to one embodiment, the method is capable of analyzing competing turndown modes of operation identified as part of the methods and processes described in connection with FIGS. 18-20. In step 730, the compiled cost data and emissions impact for each of the competing turndown modes of operation may be compared and provided as an output as part of step 731. Thus, according to how competing operating modes are compared, recommendations are provided on how the power plant should be operated during the selected turndown operating period. In turn, that recommendation includes which of the turbines should be shut down, which of the turbines should be turned down, and the turndown level at which the turbine should be operated.

  Emissions data may also be provided as part of the output of step 731, particularly if the analyzed modes of competitive operation have similar economic results. As will be appreciated, how each alternative mode affects plant emissions and, given that effect, the potential for non-compliance during the current regulatory period. Notifications may be provided, as may the economic consequences associated therewith. Specifically, the cumulative emissions of one or more power plant pollutants during the regulation period may be compared to the overall limit value acceptable during that time frame. According to a particular preferred embodiment, the step of communicating the result of the comparison is based on the emission rate calculated by averaging the cumulative emission limits over the current regulated emission period against the current regulated emission period. It may include indicating a power plant emission rate derived by averaging cumulative emission levels for the power plant over a portion. This can be done to determine how well the power plant is compared to the average emission rate that is acceptable without incurring a violation. The method is able to determine the emissions still available to the power plant during the current regulatory period, and there are sufficient levels available to support any of the proposed modes of operation. It is possible, or rather, whether the impact of emissions will unacceptably increase the probability of future regulatory violations.

  As an output, the method is capable of providing recommended measures, which are economical and otherwise between the proposed turndown and shutdown modes of operation. Advise on the advantages / disadvantages of both points. Recommendations can include reporting costs, as well as a detailed breakdown between the categories that cost them, and the assumptions made in calculating the costs. Additionally, the recommended actions can include a summary of any other considerations that may influence the decision to select the most preferred mode of operation. These can include information related to applicable emission limits and regulatory periods, and in connection therewith, what is the current cumulative emission of the power plant. This may include informing the power plant operator of the risks of violating emission thresholds, as well as any operating modes that unreasonably increase the costs associated with such violations.

  The invention may further include a unified system architecture or an integrated computing control system, which may achieve many of the functional aspects of the functional aspects set forth above. Enable and improve efficiently. Power plants, even co-owned, often operate over different markets, different government jurisdictions, and different times of day, and many types of stakeholders and decision makers who participate in their management. Exist under various types of services and other contractual commitments. Within such various settings, a single owner can control and operate multiple power plants, each of which can generate multiple power generations across overlapping markets. Have units and types. Owners can also have different criteria for assessing effective power plant operation, including, for example, unique cost models, response times, uptime, flexibility, cybersecurity, functionality. , And may include differences inherent in the manner in which different markets operate. However, as will be appreciated, most current electricity trading markets rely on various off-line generated files supplied by multiple parties and decision makers, which are used by traders, plant managers and regulators. Including those transmitted between. Given such complexity, the capacity of a power plant and / or a power generation unit within a market segment may, inter alia, extend, for example, from an individual power generation unit to a power plant, or It may not be fully understood across a multi-layered hierarchy that extends to the fleet. As such, each successive level of the electricity trading market typically hedges the performance reported by the following levels. This translates into inefficiencies and reduced revenues for owners as successive hedging increases the degree of underutilization of the system. Another aspect of the invention, as discussed below, functions to alleviate the underlying division of these problems. According to one embodiment, a system or platform is developed that performs analysis on a unified system architecture, collects and evaluates historical data, and analyzes what-if scenarios or alternative scenarios. Unified architecture enables more efficient different functions, different components such as power plant modeling, behavioral decision support tools, power plant behavior and performance prediction, and optimization according to performance goals. It is possible to According to a particular aspect, the unified architecture is for components that are local to the power plant and to the power plant, for example components hosted centrally or on cloud-based infrastructure. This can be achieved through integration with remote components. As will be appreciated, such an aspect of integration can allow for an enhanced and more accurate power plant model without affecting the consistency, effectiveness, or timelines of results. . This can include utilizing the tuned power plant model already discussed on locally and externally hosted computing systems. Given its deployment on an externally hosted infrastructure, the system architecture can be conveniently extended to handle additional sites and units.

  22-25, a scalable architecture and control system is presented that controls a fleet of a power plant in which multiple power generation units are distributed over several locations. Can be used to uphold many of the requirements associated with managing, managing, and optimizing. A hybrid local / remote architecture may be used based on context-specific or case-specific specific criteria or parameters, as provided herein. For example, certain aspects of system functionality are hosted locally, while other aspects are centrally hosted environments, such as cloud-based infrastructure, that pool data from all of the power generation units. However, it serves as a common data repository that scrubs data through cross-referenced values from common equipment, configurations, and conditions, while also supporting analytical capabilities. It may be desired by an owner or operator with a series of power plants that it can be used to: The method of choosing the right architecture for each of the different types of owners / operators can focus on the critical interests that drive the operation of the power plant, as well as the unique characteristics of the electricity market in which the plant operates. It is possible. According to particular embodiments, performance calculations may be performed locally to support closed-loop control of a particular power plant, improve cybersecurity, or provide near real-time, as provided below. It is designed to provide the response speed needed to accommodate processing. On the other hand, the system may be configured such that the data flow between the local and remote systems includes local data and model tuning parameters, where the local data and model tuning parameters are tuned power plant model. The tuned power plant model is then transferred to a centrally hosted infrastructure for the generation of and then used for analysis such as alternative scenario analysis. A remotely or centrally hosted infrastructure can be used to adapt the interaction with the common plant model according to the unique needs of different user types that need to access the common plant model. Additionally, strategies for expansion may be determined based on response times and service contracts that depend on the unique aspects of a particular market. If a faster response time for the availability of the final result is needed, the analysis process can be extended in terms of both software and hardware resources. The system architecture further supports redundancy. If any system running the analysis becomes inoperable, processing may continue on redundant nodes containing the same power plant model and historical data. The unified architecture allows applications and processes to be grouped together to improve performance and increase the range of functionality, providing both technical and commercial advantages. It has come to be realized. As will be appreciated, such advantages include convenient integration of new power plant models; separation of procedures and models; different operators while presenting data in a unique manner according to each operator's needs. Include sharing the same data in real time; convenient upgrades; and compliance with NERC-CIP restrictions for sending supervisory control.

  FIG. 22 illustrates a high level logic flow diagram or method for fleet level optimization according to certain aspects of the invention. As shown, a fleet can include multiple power generation units or assets 802, which can represent separate power generation units across multiple power plants or the power plants themselves. is there. Fleet's assets 802 may be owned by a single owner or entity and span one or more markets seeking contractual rights to generate the load sharing required by the customer grid. It is possible to compete for other such assets. Asset 802 can include multiple power generation units having the same type of configuration. In step 803, the performance data collected by the sensors on various assets of the plant can be electronically communicated to a central data repository. Then, in step 804, the measured data is reconciled, or a more accurate or more realistic indication of a performance level for each asset is determined, as described below, or Can be filtered.

  As explained in detail above, one way in which this reconciliation can be done is to compare the measured data to the corresponding data predicted by the power plant model, which As discussed, it can be configured to simulate the behavior of one of the assets. Such models can be referred to as offline models or predictive models, such models can include physics-based models, and the reconciliation process tunes the models periodically. Can be used to maintain and / or improve the accuracy with which the model represents actual behavior through simulation. That is, as discussed in detail above, the method may use the most recently collected data in step 805 to tune the power plant model. This process produces a model for each of the assets, ie a model for each of the power generation units and / or power plants, as well as a more generalized model covering aspects of the operation or fleet operation of multiple power plants. It may include tuning. The reconciliation process also ensures that data collected for resolving inconsistencies and / or for identifying anomalies, particularly data collected from assets of the same type that have similar configurations, are similar to each other. It can be accompanied by being compared between. During this process, given the collectiveness and redundancy of compiled data, significant errors can be eliminated. For example, it is possible to follow a sensor with a higher accuracy capability, or a sensor that has just been checked and has been proven to work correctly. In this way, the collected data can be cross-checked, validated, and reconciled in comparison to provide a single, consistent set of data that can be used to calculate more accurate actual fleet performance. It is supposed to be built. This set of data can then be used to tune the offline asset model, which then simulates and determines an optimized control solution for the fleet during future market periods. Can be used to enhance the competitiveness of the power plant during the power tendering procedure, for example.

  At step 806, the true performance capability of the power plant is determined from the matched performance data and the tuned model of step 805, as shown. Then, in step 807, the fleet's assets 802 may be collectively optimized subject to the selected optimization criteria. As will be appreciated, this can involve the same process already discussed in detail above. At step 808, an optimized supply curve or asset schedule may be created. This can explain the manner in which the assets are scheduled or operated, and also allow each asset to participate, for example, to meet proposed or virtual load levels for a power plant fleet. It is possible to explain the levels that can be achieved. Criteria for optimization can be chosen by the operator or owner of the asset. For example, optimization criteria may include efficiency, revenue, profitability, or some other measure.

  As shown, the subsequent step may include communicating the optimized asset schedule as part of a bid on load generation contracts for future market periods. This may include communicating the optimized asset schedule to the energy trader in step 809, which then submits a bid according to the optimized asset schedule. As will be appreciated, in step 810, the bid may be used to participate in a power system-wide power feed process, where the power system-wide power feed process causes a plurality of loads whose loads are located in the system. It is distributed between power plants and power generation units, many of which may be owned by competing owners. Bids or offers for the electricity supply process may be configured according to defined criteria, such as variable electricity generation costs or efficiencies, etc., as determined by the particular electricity distribution center of the electricity system. In step 811, the result of the power system optimization is to generate an asset schedule that reflects how various assets in the power system should be involved to meet the projected demand. Can be used. The asset schedule of step 811, which reflects the results of the overall system optimization or powering process, may then be communicated back to the owner of the asset 802, at step 812, for example, for each of the assets. The operating set point (or operating mode, among others), which may include the load at which the is operated, may be communicated to a controller that controls the operation of the asset 802. In step 813, the controller may calculate the control solution and then communicate the control solution and / or directly control the asset 802 to meet the contracted load requirements during the powering process. It is designed to let you. The fleet owner is able to adjust the manner in which one or more power plants operate according to changing conditions to optimize profitability.

  FIG. 23 illustrates data flow between a local system and a remote system, according to an alternative embodiment. As mentioned, certain functionality may be hosted locally, while other functionality may be hosted off-site in a centrally hosted environment. The method of choosing an appropriate architecture in accordance with the present invention involves determining considerations that are important drivers of the behavior of assets in the fleet. Therefore, considerations such as cybersecurity issues may require that a particular system remain local. Also, time-consuming performance calculations remain locally hosted, maintaining the necessary timelines. As illustrated in FIG. 23, the local plant control system 816 can capture sensor measurements and communicate the data to the tuning module 817, where the tuning module 817 relates to FIG. Thus, the tuning or data matching reconciliation process may be completed using performance calculations that compare actual or measured values with those predicted by the plant or asset model, as previously discussed. Through the data router 818, the model tuning parameters and reconciled data can then be communicated to a centrally hosted infrastructure, such as a remote central database 819, as shown. From there, the model tuning parameters are used to tune the offline power plant model 820 and then the offline power plant model 820 is used to optimize future fleet operation and substitute as described above. Specific scenario or “what-if” analysis to advise between possible or competing modes of operation of an asset fleet.

  The results of the analysis performed using the offline power plant model 820 can be communicated to the fleet operator via the web portal 821, as shown. Web portal 821 may provide users with customized access 822 for fleet management. Such users can include plant operators, energy traders, owners, fleet operators, engineers, as well as other stakeholders. Depending on the user interaction via web portal access, decisions regarding recommendations offered by the analysis performed using the offline power plant model 820 may be made.

  24 and 25 illustrate a schematic system configuration of a unified architecture according to certain alternative aspects of the invention. As illustrated in FIG. 25, the remote central repository and analysis component 825 can receive performance and measured operating parameters from a number of assets 802 to implement fleet level optimization. Fleet level optimization can be based on additional input data, such as the current amount of fuel stored and available at each power plant, the fuel associated with each power plant, etc. Regional prices, regional prices for electricity generated at each power plant, current weather forecasts, and differences between remotely located assets and / or outages and maintenance schedules Can be included. For example, a schedule of component overhauls for gas turbines may mean that short-term operation at higher temperatures is more economical. The process can then calculate a supply curve, which includes an optimized variable power generation cost for the fleet of the power plant. Additionally, the present invention, as shown, may allow for more automated bid preparation, in which, in at least certain circumstances, bids may be transferred directly to the system-wide power dispatching authority 826. The energy trader 809 is thereby bypassed. As illustrated in Figure 25, the results of power system optimization (via the system-wide power dispatching authority) are used to determine how various assets in the power system meet the projected demand. It is possible to create asset schedules that reflect how they should be involved. This asset schedule can reflect system-wide optimization, and can also be communicated back to the owner of the fleet of assets 802, as shown, to set operating point and The mode of operation is adapted to be communicated to a controller that controls each asset in the system.

  Thus, the method and system may be developed according to Figures 22 to 25, whereby the fleet of power plants operating in a competitive power system is associated with enhanced performance and future market periods. Optimized for bidding. Current data regarding operating conditions and parameters can be received in real time from each of the power plants in the fleet. The power plant and / or fleet model can then be tuned according to current data, with continued improvement in model accuracy and prediction range. As will be appreciated, this can be achieved through comparison between the measured performance index and the corresponding value predicted by the power plant or fleet model. In the next step, a tuned power plant model and / or a fleet level model is used, and for each of the power plants in the fleet based on competing operating modes simulated with the tuned model. It is possible to calculate the true power generation capacity. Optimization is then performed using the true plant capabilities and optimization criteria defined by the plant or fleet operator. Determining the optimized mode of operation may create an asset schedule that calculates an optimal operating point for each of the power plants in the fleet. As will be appreciated, the operating points may then be transferred to different power plants to control each according to their operating points, or alternatively, the operating points may be fed by the central power supply command authority. It can serve as the basis for bidding for submission to.

  Also, in connection with centralizing control and optimization of multiple power units, FIGS. 26 and 27 illustrate a power system 850 in which a block controller 855 includes multiple power blocks. It is used to control the 860. The power block 860 may define a fleet 861 of generating assets (“asset”), as shown. As will be appreciated, these embodiments provide another exemplary application of the optimization and control methods described above in more detail, but with optimization for fleet levels. Including expanding the perspective of. In doing so, the present invention may further offer ways to reduce certain inefficiencies that still affect modern power generation systems, especially power generation systems with multiple remote and various thermal power generation units. Is. Each of the assets may be any of the thermal power generation units discussed herein, such as, for example, gas turbines and steam turbines, and associated subcomponents such as HRSGs, intake conditioners, duct burners, and the like. Can be represented. An asset may be operable according to multiple power generation configurations depending on how the sub-components are involved. Power generation from multiple power blocks 860 may be centrally controlled by block controller 855. With respect to the system of FIG. 27, which will be discussed in more detail below, the block controller 855 can control the system according to an optimization process that includes asset and power block health, and Consider other factors that may be specific to one of the asset or power blocks 860, including power generation schedules, maintenance schedules, and location dependent variables. In addition, learning from operational data collected from similarly configured assets and power blocks, but not part of the fleet, can be utilized to further refine control strategies.

  Typically, a conventional asset controller (which is shown as “DCS” in FIG. 26) is local to the generating asset and operates substantially isolated. Due to this, such a controller cannot take into account the current health of the power block 860 and / or other assets that make up the fleet 861. As will be appreciated, this missing perspective, when considered from that perspective, leads to less than optimal power generation for the fleet 861. With continued reference to the methods and systems previously described, inter alia, in connection with FIGS. 3, 4, and 17-25, the present exemplary embodiment spans grouped assets or power blocks. It teaches a fleet level control system that enables several system-wide benefits, including enhanced power sharing strategies, cost efficiency, and improved efficiency.

  As shown, the control system is capable of interacting with the asset controller, as represented by block controller 855. The block controller 855 can also communicate with the grid 862, as well as the central power supply commanding authority or other operating authority associated with its management. Thus, for example, supply and demand information may be exchanged between the fleet 861 and the central authority. According to exemplary embodiments, supply information, such as supply bids, etc., can be based on block controller optimizations of fleet 861. The present invention can further include an optimization process that occurs between bid periods, which periodically optimizes the scheme in which the fleet 861 is configured to meet the load levels already established. Can be used to Specifically, such bid-to-bid optimization may be used to address dynamic and unexpected operating variables. Appropriate control measures for the assets of the power block 860 may be communicated by the block controller 855 to the control system in each of the power blocks 860, or more directly to the assets. According to a preferred embodiment, the implementation of the block controller 855 control solution may include allowing it to override the asset controller when certain predetermined conditions are met. . Factors affecting such overrides include variable power generation costs for each power block / asset, remaining useful component life of the hot gas path components, changing levels of demand, changing ambient conditions, and others. Can be included.

  The block controller 855 may be communicatively linked to a number of power blocks 860 of the fleet 861, as shown, as well as directly to the asset, thereby providing more data input. The control solution that can be received and is described herein is based on that data entry. The optimization procedure can consider one or more of the following inputs: health and performance degradation; power generation schedule; grid frequency; maintenance and inspection schedule; fuel availability; fuel cost; fuel use. Situation patterns and forecasts; Past problems and equipment failures; True performance capabilities; Lifting models; Startup and shutdown features; Past and present measured operating parameter data; Weather data; Cost data, etc. As discussed in more detail in connection with other embodiments, the inputs can include detailed current and historical data regarding measured operating parameters for each of the fleet 861's generating assets. . All such past and present inputs may be stored according to conventional methods, eg, in a central database, and thereby required according to any of the procedural steps described herein. Thus, it is made available in response to a question from the block controller 855.

  The cost function can be developed according to the fleet operator's choice. In accordance with the preferred embodiment, a weighted average sum of free-robustness indexes may be used to determine a preferred or optimized power sharing configuration. The fleet robustness index may include optimizations, for example, according to some factors applicable to a given demand or fleet power level. These factors can include thermal and mechanical stress; degradation or loss, including rate of degradation; cost of power generation; and / or fuel consumption. As such, the present embodiments may be used to address some ongoing issues associated with fleet control, and among other things, optimize performance across several power blocks with multiple different generation assets. To do.

  The data inputs can include those of the types already discussed herein, including those associated with computer modeling, maintenance, optimization, and model-free adaptive learning processes. For example, according to this embodiment, a computer model, transfer function, or algorithm is an asset and / or, collectively, a behavior of a power block or fleet (or a particular aspect of the behavior) of various scenarios. It can be developed and maintained as can be simulated below. Results from the simulations can include values for particular performance indicators, which represent the behavior of the asset, the power block's behavior and performance, or a prediction about fleet performance over a selected period of operation. Performance indicators may be selected due to known or developed correlations with one or more cost outcomes, and thus may be used to compare the economic aspects of each simulation. "Cost result", as used herein, can include any positive or negative economic effect associated with the operation of the fleet 861 over a selected period of operation. Thus, the cost results can include any income earned from the generation of power over that period, as well as any operating and maintenance costs incurred by the fleet. These operating and maintenance costs can include the resulting degradation to the fleet's assets, given the scenario and the resulting simulated operation from each. As will be appreciated, the data extracted from the simulation results can be used to calculate which of the alternative modes of operation for the fleet is more desirable or cost effective.

  Models for assets, blocks, or fleets include algorithms or transfer functions developed through physics-based models, adaptive or learned "model-free" process input / output correlations, or combinations thereof. Is possible. A baseline impairment or loss model can be developed that correlates the process inputs / outputs with the impairment or loss data for each asset type. Therefore, the degradation or loss data, and the associated cost results, may be computable based on the predicted values for the fleet's proposed, alternative, or competing operating mode operating parameters. , According to certain embodiments, may be differentiated by the manner in which assets and power blocks are involved, the manner in which power generation is shared across fleet assets, and other factors described herein. . As mentioned, learning from similarly configured assets can be used to inform or further refine the model used as part of this process. For example, a degradation model may be developed to calculate equipment degradation and loss that have been accrued given values for selected performance indicators. Such degradation can then be used to calculate the economic effect or cost result for each of the competing modes of operation. These economic effects are related to degradation to asset performance, wear to components, useful component life consumed (ie, a portion of the component useful life consumed during the period of operation), as well as, for example, emissions. It is possible to include measures of value such as cost taken, regulatory charges, fuel consumption, as well as other variable costs depending on power levels. As will be appreciated, degradation and consumption of useful component life for a particular asset can occur non-linearly and depend on dynamic and / or location-specific variables, so If the minimization is shared across the assets in order to minimally affect the fleet's power generation capacity and efficiency, then by distributing the fleet's power levels to minimize overall fleet degradation, a significant amount of Cost savings can be realized over time.

  Thus, multiple competing modes of operation for the fleet have been optimized or at least optimized, taking into account projected conditions for future market periods, which may include projected demand and forecasted ambient conditions. , May be selected for analysis and / or simulation to determine the preferred fleet operating mode. Each of the competing fleet operating modes can describe a unique power generation configuration for the fleet 861. Competing fleet operating modes can be developed to include a set of parameters and / or control settings that define the unique power generation configuration with which a particular fleet output level is reached. As mentioned, the fleet output level can be selected in a number of ways. First, it may be chosen to reflect an already known fleet output level, for example an output level established via a recently concluded feeding process, and the optimization process may be optimized. Can be used to determine an optimized fleet configuration, with the optimized fleet configuration satisfying a particular power level. Also, fleet power levels may be selected according to expected load levels, expected customer demand, and / or other anticipated conditions given historical generation records. Also, alternatively, the fleet output level can vary over a selected range. In this way, variable generation costs for the fleet 861 can be calculated and then used, for example, as part of a bidding procedure to inform competitive bidding preparation. Thus, the manner in which fleet power levels are defined by fleet operators may be used, at times supporting activities around preparing competitive bids, while at other times The power level can be chosen to support the advisor function, which operates to optimize fleet performance, as actual conditions can deviate from expected conditions.

  According to exemplary operation, a parameter set describing each of the competing fleet operating modes can be developed, as shown by the more detailed system of FIG. 27, and for each of the competing fleet operating modes. , Different scenarios or cases may be developed, within which different manipulable variables are variable over a selected range to determine the effect of changes on the overall behavior of the fleet. There is. Different cases for competing fleet operating modes may be configured to cover alternative schemes, where fleet output levels are shared across power block 860 and / or assets. According to another example, different cases may be selected based on the alternative configurations available for a particular one of the assets, which may include different schemes, and the different schemes for each of the assets. Will be involved. For example, some cases may involve the involvement of certain subcomponents of the asset, such as duct burners or intake conditioners, etc., and recommend that other assets operate at shutdown or turndown levels. However, it is increasing its power generation capacity. Other scenarios can explore situations where their asset mix is changed somewhat or vice versa.

  As illustrated in FIG. 27, the block controller 855 can be in communication with a data and analysis component 865, which can include a number of modules, which modules. The relevant data is collected, normalized, stored, and made available in response to questions from the block controller 855. The data recording module is capable of receiving real-time and historical data input from the monitoring system associated with the power generation asset. Also, modules related to performance monitoring may be included and one or more off-line models associated therewith may be maintained. Each of these modules is capable of functioning in substantial agreement with other embodiments discussed herein. A learning module may also be included for collection of operational data from similarly configured assets or power blocks that are not operating in fleet 861. As will be appreciated, this data can support a learning function, which provides a deeper and more complete asset behavioral understanding. Also, such data can be used to normalize the measured data collected from the fleet 861 so that performance degradation of the power generation asset can be accurately calculated, which includes fuel properties, It may include considering the effects of other variables that may affect output capacity and efficiency, such as ambient conditions.

  Fleet level optimization can be based on locality-dependent variables, as described in connection with FIGS. 24 and 25. These variables are conditions that are unique and can reflect conditions that apply to a particular asset or power block, for example, the current amount of fuel stored and available in each asset. Location specific prices for fuel for each asset; location specific market prices for electricity generated in each asset; current weather forecasts and differences between remotely located assets in the fleet. Points; and can include outage and maintenance schedules for each asset. For example, scheduled component overhauls for gas turbine assets may mean that short-term operation at higher temperatures is more economically beneficial. As shown, the data and analysis component 865 can include modules to account for these differences.

  The block controller 855 includes modules for power generation models (which can include asset models, block models, fleet models, and deterioration or loss models), optimizers, and cost functions, as shown. Can be further included. The asset, power block, and / or fleet model may be generated, tuned, and / or reconciled and maintained according to the methods already described herein. These models may be used to simulate or otherwise predict the behavior of the fleet, or selected portions thereof, over a selected period of operation, where the optimizer module is defined. A suitable scenario can be determined according to the cost function. More specifically, the results from the simulation can be used to calculate cost results for each, which include revenue, operating costs, degradation, useful component life consumed, across power blocks and / or fleet assets. , And other costs mentioned herein can be included. As will be appreciated, income may be determined via the market unit price times the planned output level. As mentioned, the cost calculation can include a degradation model or algorithm that correlates economic results to the way the asset behaves in the simulation. Performance data from simulation results can be used to determine operating costs, degradation, and other losses as previously described for the entire fleet. As will be appreciated, certain cost considerations, such as a fixed aspect of operating costs, etc., will not be significantly different between competing fleet operating modes, and as such Can be excluded from various calculations. Additionally, the simulations described herein may be configured to include an entire fleet of assets, or portions thereof, and also in predicting cost outcomes, as provided herein. It is possible to focus on the limited aspects of asset behavior found to be particularly relevant to.

  According to a particular embodiment, the cost function module can include a free-robustness index, effectively differentiating between alternative modes of operation. The free robustness index can represent the average sum of losses that occur in a power block. The robustness index can include a factor that indicates the cost associated with the useful component life consumed, which is consumed over assets such as hot gas path parts in gas turbines and compressor blades. It can be the sum of the component lifespans. For example, a generation asset scheduled to be shut down during a selected operating period according to one of the competing fleet operating modes will have a useful component life consumed at each shutdown / startup procedure. A corresponding economic loss will be incurred. On the other hand, power generation assets that are scheduled to operate at full load during the same operating period can cause losses commensurate with the time of their operation. As will be appreciated, such losses can be expected to depend on the particular thermal and mechanical loads expected given the load levels and operating parameters expected to satisfy the particular load level. Can be further calibrated to reflect ## EQU1 ## which may depend on factors such as expected ambient conditions, fuel characteristics, etc. Other economic losses may be included in the sum of fleet losses, resulting in cost results for each of the competing fleet operating modes. These can include the aggregate fuel consumption for fleet assets, as well as the economic impact of, for example, expected emission levels given simulation results.

  Once the sum of income and / or loss across the fleet has been completed for each of the simulated scenarios, the method comprises the step of calculating one or more suitable or optimized cases. Is possible. The method can then include one or more outputs related to the preferred or optimized case. For example, the preferred or optimized case may be communicated electronically to the fleet operator, such as through user interface 866. In such cases, the output of the method may be power block / asset health advisors, power sharing recommendations, shutdown planners, optimal set point control solutions for power blocks, DCS overrides, and / or expected power generation schedules. Can be included. Also, the output can include an automated control response, which can include automatically overriding one of the asset controllers. According to another alternative, the output may include generating a feed bid according to one or more of the preferred or optimized cases. As will be appreciated, the output of the method can enable fleet savings in a number of ways, as shown above the user interface 866. First, for example, a suitable power sharing arrangement may minimize, reduce, or, advantageously, share fleet degradation, which may be generated over future operating periods. And can significantly affect efficiency. Second, the advisor function may be configured to optimize the maintenance intervals, or at least enhance the maintenance intervals, using the described components, such that the degradation loss is recoverable. And irrecoverable ones are alleviated. Monitoring and predicting the rate of degradation and effectively scheduling / performing maintenance procedures such as compressor cleaning or filter cleaning ensure that the gas turbine operates most efficiently. Becomes

  Turning now to FIG. 28, another related aspect of the present invention is discussed, which illustrates a more specific example of controlling operation of multiple gas turbine engines operating as power blocks. are doing. As will be appreciated, the gas turbine engine may be located in a particular power plant or across several remote power plants. As discussed above, controlling the blocks of a gas turbine to optimize or enhance power sharing is a challenge. Current control systems do not effectively synchronize across blocks of multiple engines; instead, each of the engines is individually based on a simple sharing of the power levels collectively responsible for the power blocks. Get involved virtually. As will be appreciated, this often leads to unbalanced and inefficient rates of degradation. Therefore, there is a need for more optimal control strategies, especially across multiple gas turbines that promote more cost effective loss or degradation rates when the units are collectively considered as power blocks. There is a need for a system controller that provides an efficient power sharing strategy. For example, if a gas turbine block has several engines with the same rating, the present invention suggests that which of the units should operate at a higher power level, based on the engine's current degradation conditions. It is possible to make recommendations as to which and which should operate at a reduced level. The present invention is able to accomplish this in accordance with the aspects already discussed herein, and in particular according to those discussed with respect to Figures 24-27. As will be appreciated by those skilled in the art, the benefits of such functionality include increased gas turbine life and performance; improved life prediction (which allows for more competitive and / or risk sharing service agreements). ); Greater operational flexibility for the power block as a whole; and robust multi-purpose optimization that efficiently takes into account operational trade-offs, such as, for example, hot gas flow paths. Of useful component life, current degradation level, and rate of degradation, as well as current power generation performance such as demand, efficiency, fuel consumption, and so on.

  One way this can be achieved is according to system 900, where system 900 will be described with respect to FIG. As shown, multiple gas turbines 901 may be operated as part of a power block or “block 902”. Operating parameters 903 for each of the assets 901 may be collected and electronically communicated to the block controller 904, as discussed as part of the system above. According to a preferred embodiment, operating parameters can include rotor speed, compressor surge margin, and blade tip clearance. As will be appreciated, compressor surge margin may be calculated in terms of measured rotor speed and blade tip clearance may be measured according to any conventional method including, for example, microwave sensors. As a further input, the block controller can receive records 905 from database components such as any of those already discussed, such as rotor speed, surge margin, blade tip clearance, control settings, ambient. It is possible to record current and past operating parameter measurements, including conditional data, etc., to adaptively correlate process inputs and outputs.

  According to the preferred embodiment, the block controller 904 may be configured to operate as a model-free adaptive controller. The model-free adaptive controller can include a neural network based setup with inputs (eg, via record 905) from each of the gas turbines for demand, heat rate, and so on. As will be appreciated, model-free adaptive control is a particularly effective control method for unknown discrete-time nonlinear systems with time-varying parameters and time-varying structures. Model-free adaptive control design and analysis focuses on process inputs and outputs and "learns" predictive correlations or algorithms that explain the relationships between them. The correlation between the measured input and output of the system is controlled. Acting in this manner, the block controller 904 can derive control commands or recommendations, which can be communicated to the master control system 907 as output 906 for implementation. According to a preferred embodiment, the output 906 from the block controller 904 comprises suitable or optimized power sharing commands or recommendations. According to other embodiments, the output 906 is associated with a modulated coolant flow for a hot gas path component of the gas turbine 901 and / or a modulated IGV setting for a compressor unit of the gas turbine 901. Command or recommendation.

  The master control system 907 may be communicatively linked to the gas turbine 901 of the power block 902 and is adapted to implement the control solution given the output 906. Also, as shown, the master control system 907 can communicate such information to the block controller 904. As so configured, the control system of FIG. 28 is operable to control a number of gas turbines of power block 902, enhanced or according to a defined cost function. In an optimized manner to generate a combined load or power level (eg, a contract power level such as may be determined by the feed bidding process, for which the gas turbines are collectively responsible) ing. The control solution may include recommending a percentage of combined power levels that each of the gas turbines should contribute. In addition, the master control system 907 can include a physics-based model for controlling the gas turbine according to the optimized operating mode, as previously discussed.

  According to an exemplary embodiment, for example, clearance and surge margin data may be tracked for each of the gas turbines. If the clearance or surge margin data for any of the gas turbines is determined to exceed a predetermined threshold, then the particular turbine may be operated at a reduced load. If the gas turbine cannot be operated at reduced levels, other recommendations may be made, such as IGV settings or modulating the coolant flow to the hot gas path components. On the other hand, if one of the gas turbines is chosen to operate at a reduced level, then the optimized power generation configuration will be of the other gas turbines to make up for any shortfall. It may include recommending that one or more operate at higher / peak loads. The method is capable of selecting higher / peak load turbines based on surge margin and clearance data to collectively extend operating life while maintaining higher block power levels and higher efficiencies. In addition, it has the desired effect of balancing the current degradation level and rate of degradation during the gas turbine of the power block. As stated, the rate of performance degradation and useful component life consumption can occur non-linearly, and can depend on parameters that are variable across geographically dispersed units, so the blocks described herein. Savings can be realized by using the level of perspective and sharing the load in a manner that optimizes the cost result for the block. Therefore, real-time data determined to be extremely useful and efficient in assessing the level of performance degradation, rate of degradation, remaining component life, and true performance capability for the block's gas turbine (especially surge margin and By taking into account clearance data), power generation can be shared to optimize cost over block 902.

  Referring now to FIG. 29, another exemplary embodiment of the present invention includes systems and methods that provide more efficient and / or optimized shutdown of combined cycle power plants. As will be appreciated, during shutdown of a combined cycle plant, the controller typically reduces fuel flow to the gas turbine gradually, reducing rotor speed towards a minimum speed. ing. This minimum speed can be referred to as the "turning gear speed." The reason is that it represents the speed at which the rotor is engaged by the turning gear and thereby rotated to prevent thermal bending of the rotor during the shutdown period. Depending on the nature of the gas turbine engine, fuel flow may be stopped at about 20 percent of full speed and the turning gear engaged at about 1 percent of full speed. However, this gradual reduction in fuel flow does not provide a direct link with the reduction in rotor speed. Rather, a large unyielding change in rotor speed over the shutdown period is typical. The change in rotor speed can then cause a considerable difference in the fuel to air ratio, which is due to the fact that air intake is a function of rotor speed, while fuel flow is not. There is. Such changes can then lead to combustion temperatures, transient temperature gradients, emissions, coolant flow, as well as other significant abrupt changes. Changes in shutdown behavior can have an impact on turbine clearance and, thus, overall turbine performance and component life time.

  Therefore, there is a need for a combined cycle shutdown controller that improves plant shutdown by correcting one or more of these problems. Preferably, such a controller will control the rate of deceleration of the turbine rotor and associated components over time, minimizing uneven shutdown changes, thereby minimizing negative effects on the engine system and components. Is becoming According to a particular embodiment, a more effective controller functions to optimize rotor stress and rotor speed and torque slew rate. The shutdown controller is also capable of modifying subsystem variability such that, for example, coolant flow and wheel space temperatures are maintained at suitable levels. According to a preferred embodiment, the method is capable of controlling the rate of deceleration of the rotor and associated components over time, in a manner that reduces costs, plant losses and other negative effects. It is designed to minimize shutdown changes. According to the systems and methods already described, the control methodology can function such that the factors affecting the shutdown cost are optimized according to operator defined criteria or cost functions. One way this can be accomplished is according to process 920, where process 920 will be described with respect to FIG. As will be apparent to those skilled in the art, aspects of process 920 are based on the subject matter already discussed herein, with reference to, inter alia, the discussion associated with FIGS. 3 and 4. It will be summarized for brevity and not completely repeated.

  According to one embodiment, the shutdown procedure and / or combined cycle shutdown controller of the present invention is configured as a conventional loop forming controller. The controller of the present invention can include aspects of model-free adaptive control as well as model-based control, exactly as set forth in the appended claims. The combined cycle shutdown controller can include a target shutdown time controller and an actual shutdown time controller, and can control most if not all aspects of plant shutdown. Controllers include exhaust spread, wheel space temperature, clearance, surge margin, steam turbine and gas turbine rotor stress, gas turbine rotor deceleration rate, demand, fuel flow, current power production, grid frequency, secondary firing. , And it can receive inputs such as drum level, and so on. Based on these inputs, the shutdown controller, as described in more detail below, the time range for shutdown (e.g., rate of shutdown), slew rate of rotor deceleration, corrected coolant flow, and , Inlet guide vane profile, and / or desired generator reverse torque during shutdown can be calculated. According to particular embodiments, each of these outputs may be used to offset a potentially harmful shutdown change detected by one of the power plant sensors. The combined cycle shutdown controller is capable of providing RPM / slew rate trajectories for time profiles, and deceleration rates for appropriate current power production profiles due to shutdown operation, both for HRSG, steam turbine, and It is possible to take into account combined cycle systems, such as boilers. The shutdown controller is capable of controlling the components described herein until the turning gear speed is achieved, thus reducing the component stress while providing more optimal steam turbine and HRSG maneuverability conditions. I will provide a.

  29 is a flow chart illustrating an embodiment of a process 920 suitable for shutting down a combined cycle power plant, such as the power plant 12 described in connection with FIG. Process 920 may be implemented as computer code executable by a combined cycle shutdown controller and may be initiated after receiving a shutdown command (step 921). Shutdown commands may be received based on, for example, maintenance events, fuel change events, and the like. The process 920 can then read the current state of the plant component (step 922), which senses for any of the sensors, systems, and / or methods already described herein. Can be collected, stored, and read. The current state of plant components can include, for example, turbine rotor speed, component temperature, exhaust temperature, pressure, flow rate, clearance (ie, the distance between rotating and stationary components), vibration measurements, and so on. is there. The state of the plant adds current electricity production and costing data such as, for example, the cost of not producing electricity, the cost of electricity at market rates, green credits (eg emission credits), etc. It is possible to include it specifically.

  In the next step, costing or loss data may be retrieved by querying various systems including, for example, accounting systems, future trading systems, energy market systems, or a combination thereof (step 923). . Historical data may additionally be read (step 924). Historical data includes log data about system performance, maintenance data, fleet-wide historical data (eg, logs from other components in a plant located in various geographical locations), inspection reports, And / or may include historical cost data.

  The process may then derive an algorithm for plant shutdown degradation or loss associated with the shutdown operation (step 925). Such a derivation may include several inputs, including, for example, historical operational data associated with gas turbine systems, steam turbine systems, HRSG units, as well as any other sub-components that may be present. It can be determined using the type. Various models or algorithms may be developed according to particular embodiments to derive combined cycle shutdown losses. As discussed more fully in the discussion associated with FIG. 4, such algorithms may be based on selected operating parameters or performance indicators such as temperature, pressure, flow rate, clearance, stress, vibration, shutdown time, etc. It is possible to function to provide a sum of power plant shutdown losses based on the value of As will be appreciated, consistent with other embodiments described herein, an alternative proposed or competing shutdown mode for a combined cycle power plant is a combined cycle power plant model. Can be simulated in. That is, a combined cycle power plant model may be developed, tuned, and maintained, and then used to simulate alternative or competing shutdown modes of operation, and predictions for certain predetermined performance indicators. It is designed to derive the value to be calculated. The predicted values for the performance parameters can then be used to calculate the shutdown cost according to the derived loss algorithm.

  For example, an algorithm can be developed that correlates with the shutdown loss and the predicted thermal stress profile, which is the predicted value for a particular performance parameter given the operating parameters associated with one of the competing shutdown operating modes. Can be determined from According to particular embodiments, such losses may reflect the overall economic impact of competing shutdown modes of operation, such as degradation to hot gas path components and / or shutdown modes. It is possible to take into account the percentage of useful component life consumed, as well as any resulting performance degradation for the plant, which can be calculated from the start of shutdown to the realized time of shutdown, The realized time of shutdown can be, for example, when the turning gear speed is realized. Similarly, loss algorithms can be applied to compressor and turbine mechanical stresses; clearance between stationary and rotating parts; shutdown emissions; shutdown fuel consumption; steam turbine rotor / stator thermal stress; boiler drum pressure gradients, etc. It can be developed to determine the associated loss.

  Process 920 can then derive an enhanced or optimized shutdown mode of operation (step 926), which includes RPM / slew rate versus time profile and / or shutdown of the combined cycle power plant. It is possible to include a rate of deceleration for the current power production profile that is particularly well suited to For example, the optimized shutdown mode of operation may be determined as one that best accommodates and takes into account the operation of the steam turbine, HRSG unit, boiler, and / or other components of the combined cycle plant. According to one embodiment, the controller takes the above inputs and derives the expected conditions therefrom at various RPM / slew time parameters to minimize stress and / or optimize shutdown costs. The RPM / through curve plotted along the time axis is derived. Similarly, the rate of deceleration relative to the current power production profile may more desirably include fuel flow that improves current power production based on shaft deceleration (when compared to other proposed shutdown modes). It is possible. According to other embodiments, a cost function may be defined that derives and selects another more preferred or optimized shutdown mode of operation. As will be appreciated, for example, a stress and / or loss minimizing scenario for a gas turbine, a stress and / or loss minimizing scenario for a steam turbine, a stress and / or loss for a HRSG A scenario or a combination of them can be derived. Depending on how the cost function is defined, further optimized shutdown modes of operation include cost, plant emissions, fuel consumption, and / or unlimited power production during the shutdown period. Can be determined based on criteria, such as combinations thereof.

  The combined cycle shutdown controller can further include a control system for shutting down the power plant according to the optimized shutdown mode. According to a preferred embodiment, the control system can include a physics-based modeler or a model-based controller, which then derives the control solution subject to an optimized shutdown mode. . The model-based controller is capable of deriving control inputs and settings for controlling actuators and control devices, allowing the combined cycle power plant to operate in a shutdown period according to a preferred or optimized shutdown operating mode. It can be operated in the meantime. For example, the shutdown controller may operate the fuel valve to match the desired fuel flow rate while additionally controlling the steam valve of the steam turbine to control the steam turbine shutdown while controlling the gas turbine. It controls the exhaust inlet guide vanes and controls the exhaust flow into the HRSG. By combining the controls of the various components at about the same time with each other, the shutdown for the plant can be improved and adapted to the desired scenario.

  Although the present invention has been described with respect to what is presently considered to be the most practical and preferred embodiments, the present invention should not be limited to the disclosed embodiments, by contrast. It is to be understood that it is intended to cover various modifications and equivalent constructions included within the spirit and scope of the appended claims.

10 power system 12 power plant 12a power plant 12b power plant 14 transmission line 16 load 18 storage device 20 communication network 21 connection 22 plant controller 23 load controller 24 power supply command authority 25 power system controller 26 data resource 30 gas turbine system 31 component controller 32 Compressor 33 Ambient environment 34 Combustor 36 Turbine 39 Operator 40 Intake duct 41 Inlet guide vane 42 Exhaust duct 44 Generator 46 Sensor 47 Power plant actuator 49 Power plant component 50 Steam turbine system 51 Intake conditioning system 52 HRSG Duct firing system 53 Steam turbine 55 High temperature exhaust gas 6 Gas turbine model 61 Intake conditioning system model 62 Steam turbine model 63 Economic model 64 Optimizer 65 Refrigeration system 70 Filter 71 Neural network 72 Actual data 73 Predicted data 74 Optimized set point 75 Power plant model 80 Computer system 81 Display 82 Processor 83 User input device 84 Memory 110 Control module 111 Control model 112 Performance model 113 LCC model 114 Enhancement module 115 Control information 120 Offer curve module 121 Ambient condition forecast 122 Market condition forecast 123 Database 124 Offline model 125 Offer curve 126 model Tuning module 128 Target load 16 9 method 170 control section 171 step 172 step 173 step 174 step 175 step 180 offer curve section 181 step 182 step 183 step 184 step 185 step 186 step 200 plant optimization system 202 monitoring and control instrument 204 monitoring and control instrument 208 plant controller 210 Data collection module, data collection model 212 Database / Historian 214 Heat balance module 216 Performance module 218 Optimizer module 220 Advisory output 222 Closed feedback loop control output 230 Real-time thermal power plant optimization system 231 Server system 234 Client system System 236 database server 239 database 250 method 300 system 301 system 302 power plant 303 controller 303a controller 303b controller 305 first model, primary model 306 second model, predictive model 307 interface 308 current performance parameter 309 other variables 310 Control profile input 313 Operator specific scenario 314 Predicted turbine operating characteristics 320 Method 325 Step 330 Step 335 Step 340 Step 345 Step 400 Method 405 Decision Step 410 Step 415 Step 420 Decision Step 425 Step 430 Step 435 Step 440 Step 445 Step 500 Floda Agram 501 power plant 502 power plant model 503 tuning module 505 plant controller 507 tuned power plant model 509 plant operator module 510 optimizer 511 sensor 515 feedback 516 performance goal 517 second parameter set 519 simulated behavior 521 optimization Mode of Operation 600 Algorithm 601 Maximum 602 Threshold, Step 603 Algorithm 604 Step 606 Step 608 Step 609 Exemplary Permutation Matrix 610 Algorithm, Step 611 Maximum Chilling Capability, Step 612 Related Algorithms, Operator 613 Dew Point, Step 614 Step 616 Step 618 Step 620 Algorithm, Step 621 Upper Threshold Level, Step 622 Step 630 Chart 631 Chart 639 Sequence 640 Initial State 700 Flow Diagram 701 Turndown Data 702 Shutdown Data 703 Common Data 710 Turndown Analyzer 711 Step 712 Step 713 Step 714 Step 716 Step 717 step 718 step 719 shutdown analyzer 720 step 721 step 722 step 723 step 724 step 726 step 727 step 728 step 729 step 730 step 731 step, output 802 asset 803 step 804 step 805 step 806 step 8 7 Steps 808 Steps 809 Steps, Energy Traders 810 Steps 811 Steps 812 Steps 813 Steps 816 Local Plant Control Systems 817 Tuning Modules 818 Data Routers 819 Remote Central Databases 820 Offline Power Plant Models 821 Web Portals 822 Customized Access 825 Remote Central Repositories And analysis component 826 System-wide power supply authority 850 Power system 855 Block controller 860 Power block 861 Fleet 862 Grid 865 Analysis component 866 User interface 900 System 901 Gas turbine, Asset 902 Power block 903 Operation parameter 904 block controller 905 recording 906 output 907 master control system 920 process 921 step 922 step 923 step 924 step 925 step 926 step 927 step

Claims (14)

A control method for optimizing the operation of the power plant fleet, the power plant fleet comprises for generating electricity sold in the market, a plurality of assets position to a remote, the generator plant fleet includes a plurality of operating configurations that will be differentiated plurality of assets by how related or Azukasuru, said plurality of assets can be described by operating parameters relating to aspects of the physical operation (903) It has multiple operation modes and the control method is
Sensing and collecting a measurement of the operational parameter (903) for the operation of each of the plurality of assets;
Comprising the steps of tuning the asset model to construct the asset model tuned for each of the plurality of assets, the step of tuning includes data reconciliation process, in the data reconciliation process, a predetermined operation The measured value for the parameter (903) is compared to the predicted value for the predetermined operating parameter (903) to determine the difference between them, and the tuning of the asset model is performed. The step is based on the difference, and
Simulating a proposed operating configuration of the power plant fleet using the tuned asset model, the simulation of the proposed operating configuration comprising:
Defining a predetermined operating period,
Selecting the proposed operating configuration, and
Performing a simulation run on each of the proposed operating configurations using the tuned asset model, whereby the operation of the power plant fleet during the predetermined operating period is simulated. And including steps, and
A step of obtaining a simulation result from each of the simulation run, each of said simulation results, including the predicted value of the performance index, and a step seen including,
The control method further comprises determining, from the proposed operating configuration, an optimized operating configuration for the predetermined operating period for the power plant fleet, wherein the determining the optimized operating configuration is economic. A cost-effective optimization process, the economical optimization process comprising:
Defining a cost function,
Evaluating each of the simulation results according to the cost function and determining from them an optimized simulation run;
Specifying any of the proposed operating configurations as the optimized operating configuration of the power plant fleet, even if it corresponds to the optimized simulation run;
Including,
The tuned asset model for each of the plurality of assets includes a common tuned asset model, the common tuned asset model being local to a location of the plurality of assets to which it corresponds. And control methods that are operably duplicated and maintained both at remote locations . The performance indicators include desired criteria for evaluating the performance of the power plant fleet over the predetermined period of operation,
The cost function, the predicted value for said performance index includes an algorithm that correlates the expected economic consequences, control method according to claim 1, wherein. The control method according to claim 2 , wherein the performance index includes at least one of fuel consumption and a heat rate for the power plant fleet over the predetermined operation period. Each of the previous SL multiple assets, some of the power plant in the power plant fleet (12,12a, 12b, 302,501) thermal power unit being distributed between, and, of the power plant fleet One of the power plants (12, 12a, 12b, 302, 501) therein, said power plant (12, 12a, 12b, 302, 501) each comprising a plurality of thermal power generation units. 3. The control method described in 3 . The local replica of the common tuned asset model is designated as the local tuned asset model, and the remote replica of the common tuned asset model is designated as the remote tuned asset model Has been done,
The locally maintained common tuned asset model comprises a local tuned asset model and the remotely maintained common tuned asset model comprises a remote tuned asset model. ,
It said remote tuning asset model of said plurality of assets is in position on a common remote location, the control method according to any one of claims 1 to 4. The control method of claim 5 , wherein the common remote location comprises a central data repository. The control method of claim 5 , wherein the common remote location comprises a cloud-hosted system. The control method further comprises the step of periodically synchronizing the local tuned asset model with the remote tuned asset model to configure the common tuned asset model. 5. The control method described in 5 . The local tuned asset model for each of the plurality of assets is operably linked to a control system for the plurality of assets to which it corresponds and supports real-time processing associated therewith. 6. The method of claim 5 , wherein the real-time processing comprises a locally performed performance calculation associated with closed-loop control of the plurality of assets. The local tuned asset model and the remote tuned asset model for each of the plurality of assets are operably configured to provide redundant support for at least one type functionality. The control method according to claim 5 . The control method is
Electronically sending the measurements of the operating parameters (903) collected in each of the plurality of assets to the common remote location;
Validating a data set for each of the plurality of assets through inter-comparison of the measurements between the assets of the plurality of assets that include the same type and similar behavioral configuration,
The control method of claim 5 , wherein the step of tuning the asset model comprises using the measurements from the validated data set. The step of tuning is completed at the common remote location using the remote tuned asset model,
The control method uses the local tuning parameters derived from the tuning of the remote tuned asset model to synchronize the local tuned asset model with the remote tuned asset model. The control method of claim 11 , further comprising the step of electronically communicating to the tuned asset model of. Wherein said step of tuning is completed before Symbol using local tuned asset model,
The control method comprises: tuning parameters derived from the tuning of the local tuned asset model to the remote tuned asset model to synchronize the local tuned asset model with the remote tuned asset model. The control method of claim 5 , further comprising the step of electronically communicating to the tuned asset model of. The step of selecting the proposed operating configuration for the power plant fleet comprises defining competing operating modes, and then defining for each of the competing operating modes a plurality of cases associated therewith,
The control method further comprises generating a proposed parameter set for each of the plurality of cases defined for the competing operating modes, the proposed parameter set comprising the tuned power plant model ( The control method according to claim 5 , further comprising input data relating to (507), said input data defining a value relating to a predetermined variable during said predetermined operation period.